key: cord-1038897-cr6wlql2
authors: Autio, Timo J.; Koskenkorva, Timo; Leino, Tuomo K.; Koivunen, Petri; Alho, Olli‐Pekka
title: Longitudinal analysis of inflammatory biomarkers during acute rhinosinusitis
date: 2016-10-18
journal: Laryngoscope
DOI: 10.1002/lary.26344
sha: 51e5487e7da1989ecf0dcf61fb5d95ff0ce14ed5
doc_id: 1038897
cord_uid: cr6wlql2

OBJECTIVE: To illuminate the pathophysiology of acute rhinosinusitis (ARS) with sequential monitoring of inflammatory biomarkers during an ARS episode and to clarify their diagnostic usability in bacterial ARS. STUDY DESIGN: Inception cohort study with 50 conscripts with ARS. METHODS: We collected peripheral blood high‐sensitive C‐reactive protein (hs‐CRP), white blood cell (WBC), procalcitonin, and nasal nitric oxide (nNO) counts at 2 to 3 and 9 to 10 days of symptoms during an ARS episode. We simultaneously gathered various clinical parameters and microbiological samples. Bacterial ARS was confirmed with a positive culture of sinus aspirate. RESULTS: Reciprocal correlations and a significant change in biomarker levels between the two visits suggest that ARS involves a local and systemic inflammatory response that was strongest at 2 to 3 days. High‐sensitive CRP and nNO reflected responses best (52% had increased CRP levels at 2–3 days; 66% had decreased nNO levels). White blood cell and procalcitonin counts rarely exceeded the reference range. Increased local and systemic inflammatory response were linked to multiple, adenoviral, or influenza A viral etiology or the detection of bacterial ARS. Local response correlated with imaging findings of wide paranasal sinus involvement and ostiomeatal complex occlusion. At 9 to 10 days, elevated (≥ 11 mg/L) and moderately elevated (≥ 49 mg/L) hs‐CRP predicted bacterial ARS well (likelihood ratio [LR]+ 3.3 and LR+ 15.8, respectively), but the sensitivity for both findings remained low. CONCLUSION: Acute rhinosinusitis (particularly bacterial ARS) involves a local and systemic inflammatory response that is strongest at the beginning of symptoms. Elevated hs‐CRP supports the diagnosis of bacterial ARS. LEVEL OF EVIDENCE: 4. Laryngoscope, 2016 127:E55–E61, 2017

Acute rhinosinusitis (ARS) is a complex entity with an exact pathogenesis that still is partly obscure. In most cases, ARS involves both the nasal passages and the paranasal sinuses, 1-3 and both viruses and bacteria are involved in the pathogenesis. [4] [5] [6] First, viral infection and middle meatal pathogenic bacterial coinfection can modify the disease process, and the end result may be bacterial secondary infection of the sinuses. [7] [8] [9] The key pathogenic process of ARS is the inflammatory reaction of the nasal and paranasal mucosa, [10] [11] [12] [13] which may theoretically be linked to the microbiological etiology, the spread of the disease to the paranasal sinuses, and to symptoms.

In theory, biomarkers that parallel the local and systemic inflammatory reaction could shed light on the pathophysiology and development of bacterial ARS. Peripheral blood leucocytes (white blood cell count [WBC] ) and C-reactive protein (CRP) as well as nasal nitric oxide (nNO) have been suggested as possible biomarkers of disease activity in ARS. 14 In addition, a proinflammatory biomarker-procalcitonin (PCT)-has been proposed as a pertinent marker of bacterial infection. 15 Longitudinal studies with wild and experimental colds have revealed that about half of infected subjects present a mild increase in CRP, but these studies have not examined whether the subjects developed bacterial ARS. 16, 17 Others have shown in cross-sectional studies that microbiologically confirmed bacterial ARS is associated with elevated CRP and WBC levels at the time of the diagnosis, but the studies have not followed up these biomarkers systematically from the beginning of the ARS episode. 18, 19 We hypothesized that systematic sequential monitoring of these biomarkers in the same patients during a single ARS episode with microbiological confirmation of bacterial ARS would provide a further view of the pathophysiology of ARS.

We conducted an inception cohort study among conscripts during a single episode of ARS and took sequential total WBC, high-sensitive CRP (hs-CRP), PCT, and nNO counts. Our aim was to 1) investigate the biomarker levels during the ARS episode and particularly in relation to the development of culture-proven bacterial ARS; 2) look at the relationship between the biomarker levels and microbiological and clinical findings; and 3) determine the diagnostic accuracy of these biomarkers in bacterial ARS.

We conducted a prospective inception cohort study among conscripts completing their compulsory military service in the Kainuu Brigade in Northern Finland. All of the patients provided written informed consent, and the study protocol was approved by the ethics committee of the Northern Ostrobothnia Hospital District, Finland.

We have used this same study population to study the microbiology, 9 clinical diagnostics, 20 and imaging findings 21 of ARS. Briefly, the patients were recruited and examined at the health center of the Kajaani Garrison, Finland. The participants were enrolled between February 1, 2012, and April 15, 2012, from consecutive conscripts who sought medical care due to ARS symptoms. The participants had to have ARS determined by the following inclusion criteria: acute onset within 4 days, presence of nasal symptoms (blockage or discharge), 10 and abnormal nasal findings (mucosal edema or secretion). Patients excluded had concomitant infection necessitating antimicrobial treatment, respiratory infection, or antimicrobial treatment within 3 weeks preceding the first visit; nasal allergy or asthma requiring medication; chronic nasal symptoms or polyps; or prior nasal or sinus surgery.

At enrollment, we collected background information and examined the patients. A follow-up visit was scheduled for the workday closest to the 10th day of symptoms. The recorded symptoms, imaging examinations, and microbiologic specimen and analyses have been described earlier in detail. 9, 20, 21 What follows is a brief description of these data.

Recording of Symptoms. The participants recorded in a symptom diary the presence and severity (0 5 none to 10 5 worst possible) of acute symptoms daily for 10 days. 10 The following symptoms were recorded: nasal blockage, clear nasal discharge, purulent nasal discharge, postnasal discharge, reduction/loss of smell, and facial pain/pressure. The diary started from the onset of symptoms. We summed up the grades for each symptom to obtain a total symptom score (0-60) for each day. 9, 20 Cone-Beam Computed Tomography of the Paranasal Sinuses. A paranasal cone-beam computed tomography (CBCT) scan was taken at enrollment and at the follow-up visit. Coronal, axial, and sagittal slices of the maxillary, ethmoid, and sphenoid sinuses and the ostiomeatal complex (OMC) were visualized. Both sides of the paranasal sinuses were graded separately as follows 22 : grade 0 (normal), 3 mm mucosal thickening; grade 1 (minor abnormality), 4 mm mucosal thickening at any point in the sinus 23 ; and grade 2 (major abnormality), air-fluid level, gas bubbles, or total opacification. The ostiomeatal complex (OMC) was graded as 0 (open) or 2 (occluded). We determined the total CBCT score (0-16) by summing up the grades of each sinus and the OMC on both sides.

Microbiological Specimen and Analyses. We took samples for multiplex respiratory virus polymerase chain reaction from the nasopharynx at both visits and the sinus aspirates at the follow-up visit. Searches were conducted for 12 respiratory viruses. We took the middle meatus bacterial samples at both visits. If the paranasal CBCT scan at the follow-up visit showed total opacification, air-liquid level, or gas bubbles of either maxillary sinus, the patient underwent maxillary sinus aspiration under topical anesthesia. The aspirate was cultured to identify culture-proven bacterial ARS. 20 

At enrollment and the follow-up visit, we took blood samples to determine the WBC count and the concentrations of hs-CRP and PCT, which reflect systemic inflammation. In addition, we measured the level of nNO, which mirrors local inflammation.

We analyzed the WBC count using an automatic impedance and flow cytometry analyzer (Celltac MEK-6420, Nihon Kohden Ltd, Tokyo, Japan), hs-CRP with a photometric and immunochemical method (Roche/Hitachi Modular P800 chemistry analyzer, Hitachi High-Technologies Corporation, Tokyo, Japan), and PCT using a chemiluminescence immunoassay (Advia Centaur XP, Siemens Healthcare GmbH, Erlangen, Germany). The minimum detectable levels of CRP and PCT were 0.05 mg/L and 0.05 lg/L, respectively. The levels of nNO were measured with an electrochemical device (Niox Mino Nasal, Aerocrine, Solna, Sweden) using a firm nasal olive and direct sampling from both nasal cavities with a 5-mL/ minute airstream generated by the analyzer. Pursed-lip breathing was performed by the patient during the measurement to maintain velum closure and to prevent nNO leakage through the velopharyngeal space. 24, 25 The minimum concentration (parts per billion) of the right and left nasal counts was used for each subject.

For descriptive data, we calculated the median and interquartile range. To test the change in biomarker levels over time, we used the paired Wilcoxon signed rank test. To make the results comparable, the hs-CRP, WBC, and PCT levels were categorized similarly to earlier research. 18, 19, 26 We used a nonparametric Mann-Whitney U test and a Kruskal-Wallis test to compare continuous variables. Correlation coefficients were calculated using Spearman rank correlation coefficients (r s ). All of the viral and bacterial samples were analyzed per patient.

Associations between culture-proven bacterial ARS and raised hs-CRP and WBC counts were studied further by cross-tabulation and calculation of sensitivity, specificity, 95% confidence intervals, and positive and negative likelihood ratios (LR1 and LR2).

The opportunity to participate was offered to 66 patients. Of those, 51 consented and were enrolled. One patient was excluded because of pneumonia. Details of the demographic characteristics and the ARS episode of the remaining 50 patients are shown in Table I . Those seven patients (14%) who did not return the symptom diary were excluded from the analyses of symptoms. Twenty patients (40%) underwent maxillary sinus aspiration. The culture of the aspirate was positive for nontypeable Haemophilus influenzae (NTHi) in eight (16%) cases, which thus had culture-proven bacterial ARS.

On average, hs-CRP and PCT levels decreased and nNO levels increased between the two visits, indicating both local and systematic inflammatory reaction (Table  II) . At 2 to 3 days, hs-CRP counts had a moderate correlation with WBC (r s 5 0.46, P 5 .001), PCT (r s 5 0.48, P < .001), and nNO counts (r s 5 20.42, P 5 .002). Age, sex, and smoking status were not associated with any of the inflammatory biomarkers (data not shown). At 2 to 3 days, 36% of the subjects had mildly elevated hs-CRP levels and 16% had moderately elevated hs-CRP levels. At 9 to 10 days, these proportions had declined to 18% and 8%, respectively. In 66% of the subjects, the nNO increased between the two visits. At both visits, a significantly larger proportion of subjects who eventually developed bacterial ARS had elevated hs-CRP levels as compared to subjects who did not (Fig. 1A) . However, even among subjects with bacterial ARS, the median hs-CRP level was markedly higher at 2 to 3 days than at 9 to 10 days. At both visits, subjects with bacterial ARS had significantly lower median levels of nNO than those without (Fig. 1B) .

WBC counts showed similar results as hs-CRP, but a substantially lower proportion of subjects had elevated WBC counts (Table II ). In the overwhelming majority of subjects, procalcitonin levels did not increase over the reference range during ARS.

At 2 to 3 days, the finding of multiple viruses, adenovirus or influenza A virus in the nasopharynx was associated significantly with raised hs-CRP values or decreased nNO levels (Table III) . The spreading of the disease and ostiomeatal occlusion shown in CBCT scans associated with lower nNO levels but not with higher hs-CRP levels. At 9 to 10 days, the presence of NTHi in the sinus aspirate associated both with raised hs-CRP values and decreased nNO values.

At 2 days, among all subjects the main symptom score correlated with neither hs-CRP nor nNO levels (r s 5 20.04, P 5 .79 and r s 5 0.04, P 5 .79, respectively). However, among subjects with bacterial ARS, a strong correlation was found between the hs-CRP level and the main symptom score (r s 5 0.74, P 5 .04).

At 9 to 10 days, elevated hs-CRP ( 11 mg/L) predicted bacterial ARS fairly well (LR1 3.3) and moderately elevated hs-CRP ( 49 mg/L) very well (LR1 15.8) ( Table IV) . The sensitivity figures for both of these findings, however, were low.

We conducted a prospective inception cohort study among 50 conscripts with ARS and measured various inflammatory biomarkers at 2 to 3 days of symptoms and again after 9 to 10 days. Our aim was to evaluate the pathophysiology and development of bacterial ARS and to clarify the diagnostic usability of the biomarkers. We found a significant change in hs-CRP, PCT, and nNO levels between the two visits, indicating the presence of a local and systemic inflammatory response during ARS. The significant reciprocal correlations between the biomarkers further support this finding. High-sensitive CRP and nNO levels reflected these inflammatory responses best. About half of the subjects had mildly or moderately increased hs-CRP levels at 2 to 3 days. White blood cell and PCT counts exceeded the reference range only rarely.

According to the hs-CRP and nNO values, the local and systemic inflammatory response were clearly greatest at the beginning of symptoms. Compared with nonbacterial ARS, bacterial ARS encompassed a more powerful local and systemic inflammatory reaction that was strongest at 2 to 3 days. Overall, we found the increased local and systemic inflammatory response to be linked to multiple, adenoviral, or influenza A viral etiology. Local response but not the systemic response correlated strongly with the spread of the disease to the paranasal sinuses or to occlusion of the ostiomeatal complex in a CBCT scan. In bacterial ARS, we found that the systemic inflammatory response correlated powerfully with the symptom score at the beginning of the ARS, which may indicate that the stronger symptoms in bacterial ARS could partly be the result of the enhanced inflammatory response.

The facts that the local and systemic inflammatory responses were stronger in bacterial ARS than in nonbacterial ARS, and that these reactions were strongest at the beginning of respiratory symptoms, indicates that bacterial ARS starts to develop during an early phase of symptoms. This is in agreement with our earlier finding that there is a gradually increasing discrepancy in CBCT and symptom scores between bacterial and nonbacterial ARS that begins at the start of the symptoms. 21 Furthermore, we have shown earlier that the presence of pathogenic bacteria in the nasal middle meatus (coinfection) at the beginning of viral ARS modifies the symptoms and is one factor that may lead to the finding of the same bacteria in the maxillary sinus. 9 However, here the middle meatal NTHi was not associated with raised CRP or decreased nNO. This raises the question of whether the pathogenic bacteria were already present in the maxillary sinus at the beginning of symptoms. We do not know this because we performed the maxillary sinus aspirations and bacterial cultures at the follow-up visit. This would also argue against the generally presented hypothesis that bacterial ARS follows an antecedent viral ARS. 11 At 9 to 10 days, elevated hs-CRP ( 11 mg/L) predicted bacterial ARS relatively well (LR1 3.3) and predicted moderately elevated hs-CRP ( 49 mg/L) very well (LR1 15.8), but the sensitivity figures for these findings were low. We have earlier examined the respective figures for various symptoms and clinical findings in this same material. 20 Compared to these diagnostic measures, the finding of elevated CRP ( 11 mg/L) proved to be better at predicting bacterial ARS than symptoms; it paralleled the findings of moderate or profuse amount of nasal secretion and cervical adenopathy but was poorer than the finding of nasal secretion in the posterior pharynx (supplementary material). Moderately elevated CRP ( 49 mg/L) accurately predicted bacterial ARS but only was present in four patients. Because of the relatively high negative likelihood ratios, the absence of elevated CRP could not be used to rule out bacterial ARS. Thus, CRP cannot be recommended for routine screening of bacterial ARS. Because the WBC and PCT counts rarely exceeded the reference values, their diagnostic value was low.

Our findings that about half of the cases had raised hs-CRP values at the beginning of ARS agree with those of Whicher et al. and Melbye et al. 16, 17 We do not know which proportions of patients suffered from bacterial ARS in these two studies because the researchers did not look for the presence of bacterial ARS. Here, the nasopharyngeal findings of multiple viruses, adenovirus, or influenza A virus was associated with the inflammatory response. Earlier, Ruuskanen 18, 19 Patients with radiologic sinusitis have been found to have lower maxillary sinus NO levels than control subjects. 28 One proposed mechanism for this is blockage of the sinus ostia, which reduces NO levels in nasally exhaled air. 29 In support of this, we found that bacterial ARS involved lower nNO levels than nonbacterial ARS and that ostiomeatal occlusion confirmed by imaging associated with low nNO levels.

To minimize selection bias, we recruited consecutive patients. We were not able to verify viral etiology in all the subjects, similar to prior research on natural colds. 30 We did not perform maxillary sinus aspiration to every patient, but only to those with CBCT findings that supported bacterial ARS. This might have caused information bias. However, CT is considered to be the best imaging modality for evaluating paranasal sinus disease. 10 We defined bacterial ARS to include only the maxillary sinuses for which sinus aspirate is possible. Due to available resources, the study group was relatively small. Despite this, we found significant associations between hs-CRP and nNO levels and various clinical parameters. We wanted to increase the possibility of encountering frequent bacterial ARS; therefore, conscripts were recruited. The amount of patients who get bacterial sinusitis during ARS in primary care usually is estimated to be 2% 10, 11 We would have needed 400 such patients to get the eight cases of bacterial ARS that we had in this study.

Concerning the generalizability of our results, several issues require recognition. Due to the military environment, most of the conscripts were men. Smoking was common but did not correlate with the biomarker levels. The only bacteria that we found from the maxillary sinuses were NTHi. This is probably explained by the fact that NTHi earlier had been reported to be a major nasal and paranasal pathogen among Finnish conscripts. 31, 32 Overall, the present finding of ARS involving an inflammatory response warrants reconfirmation among other patients and bacteria.

Acute rhinosinusitis (particularly bacterial ARS) involves a local and systemic inflammatory response that is strongest at the beginning of symptoms. Elevated hs-CRP supports the diagnosis of bacterial ARS. 

Computed tomographic study of the common cold

Paranasal sinus findings in children during respiratory infection evaluated with magnetic resonance imaging

Subjects with allergic rhinitis show signs of more severely impaired paranasal sinus functioning during viral colds than nonallergic subjects

Acute community-acquired bacterial sinusitis: the value of antimicrobial treatment and the natural history

The common cold

Staphylococcus aureus is a major pathogen in acute bacterial rhinosinusitis: a meta-analysis

Clinical practice guideline: adult sinusitis

Occurrence of asymptomatic sinusitis in common cold and other acute ENT-infections

The role of microbes in the pathogenesis of acute rhinosinusitis in young adults

European position paper on rhinosinusitis and nasal polyps 2012

IDSA clinical practice guideline for acute bacterial rhinosinusitis in children and adults

Canadian clinical practice guidelines for acute and chronic rhinosinusitis

Clinical practice guideline (update): adult sinusitis

Rhinosinusitis: developing guidance for clinical trials

Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections

Acute phase response of serum amyloid A protein and C reactive protein to the common cold and influenza

The course of C-reactive protein response in untreated upper respiratory tract infection

Symptoms and signs in culture-proven acute maxillary sinusitis in a general practice population

Do simple laboratory tests help in etiologic diagnosis in acute maxillary sinusitis?

Diagnostic accuracy of history and physical examination in bacterial acute rhinosinusitis

Imaging follow-up study of acute rhinosinusitis

Quantification for staging sinusitis. The Staging and Therapy Group

Paranasal sinuses on MR images of the brain: significance of mucosal thickening

ATS/ERS recommendations for standardized procedures for the online and offline measurement of exhaled lower respiratory nitric oxide and nasal nitric oxide

Exhaled and nasal nitric oxide measurements: recommendations. The European Respiratory Society Task Force

C-reactive protein in respiratory virus infections

Reduced nitric oxide in sinus epithelium of patients with radiologic maxillary sinusitis and sepsis

Update on the use of nitric oxide as a noninvasive measure of airways inflammation

The relationship between atopic status and IL-10 nasal lavage levels in the acute and persistent inflammatory response to upper respiratory tract infection

Bacteriological findings of acute maxillary sinusitis in young adults

Comparison of the nasal bacterial floras in two groups of healthy subjects and in patients with acute maxillary sinusitis