INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although it is known that oral mucosal fluids play a vital role in maintaining the health of the oropharynx and teeth through their enzymatic and immunoglobulin content, it is not known how much of a protective effect may be directly related to mechanical clearance of potential pulmonary or oropharyngeal pathogens. In normal hosts, more than 90% of effective clearance of gram-negative bacilli (GNB) from the oropharynx appears to be due to effective salivary flow and swallowing (1, 2). In the elderly debilitated patient, both flow and swallowing are frequently abnormal. This reduction in mechanical clearance of potential pulmonary or oropharyngeal pathogens may be the first step in the path that leads sequentially from oropharyngeal colonization to pneumonia.

Impairment in clearance has a potential twofold effect on oropharyngeal bacterial colonization, promoting both bacterial growth and invasion. We speculate that diminished clearance would allow pathogenic organisms increased time in the mouth for proliferation and that this change in mucosal bacterial burden promotes oropharyngeal inflammation, i.e., the influx of neutrophils and release of mediators of inflammation. These events could then initiate the well-described alterations in salivary enzyme concentrations and epithelial cell surface integrity that lead to bacterial adhesion, colonization, and infection (3). Although this putative correlation between decreased clearance and increased pathogenic oropharyngeal colonization is intuitively appealing, there are currently no data available to prove its validity by simultaneously quantitatively assessing physical clearance and organism burden.

In this investigation we examine the relationship between colonization and oral clearance in the institutionalized elderly. This population was chosen because (as an experimental model) they have both an increased incidence of reduced salivation and gram-negative colonization (9 to 60% incidence with increasing colonization with increasing functional impairment) (8). We set out to assess whether they demonstrated a decrease in physical clearance in order to investigate if and how this defect is causal in colonization with pathogenic organisms.

A prospective cross-sectional analysis was performed in 75 persons who were either hospitalized in a geriatrics ward or were elderly nursing home residents. We assessed clinical parameters identified as risk factors for xerostomia and examined and related these to oral clearance and colonization status. Oropharyngeal mucosal inflammation was monitored with salivary epithelial and inflammatory cell counts and elastase concentrations.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Seventy-five subjects were enrolled in the study. These included patients institutionalized in the Long Island State Veteran's Home, a nursing facility affiliated with University Medical Center, and geriatric patients hospitalized at University Medical Center Stony Brook. All patients signed informed consent as approved by the Human Investigations Committee.

Oral Clearance Studies

A modified protocol of LaForce and coworkers (2) was used to label the oropharynx as previously described (9). Briefly, we used a DeVilbiss model 152 atomizer (Somerset, PA) and administered ^99mTc pertechnetate bound to 21 mg human serum albumin (HSA). Three sprays were administered to the patient's mouth, one directed to the posterior wall of the pharynx, the two others directed to the buccal surface of each cheek. The time of the spraying was defined as time zero. Sixty-second recordings of oropharyngeal radioactivity were performed at 5-min intervals over 2 h with a scaler ratemeter (model 2200; Ludlum Measurements Inc., Sweetmeter, TX). The ratemeter was used in conjunction with a collimator designed to focus attention on oropharyngeal activity. We have previously shown that that this ratemeter's accuracy is similar to that of a gamma camera with an excellent correlation of data (r = 0.967, p < 0.001) (9). Day-to-day variability was assessed by serial studies 1 wk apart in eight subjects (Figure 1). There were no significant changes in percent retention in the two studies.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Percent retention of the radiolabel over 120 min in patients (n = 8) who were studied twice with a 1-wk interval between studies. There were no significant differences between these curves. The error bars represent 1.96% × SEM (95% CI).

Microbiologic Data

Oropharyngeal cultures were obtained on the day of the oral clearance study by rubbing the buccal mucosa with a sterile cotton swab which was then placed in transport medium. All cultures were inoculated within 30 min of collection on blood, MacConkey's and chocolate agar and were aerobically incubated at 35° C.

Functional and Nutritional Assessment

Functional status was assessed by Barthel index, a functional assessment tool which measures a wide range of activities (10). Mental function was qualified using Folstein's Mini-Mental State Exam (MMSE), which measures attention span and constructional ability (11). Nutritional status was assessed using serum albumin.

Salivary Cell Populations, Elastase Activity

Saliva was obtained by repeated expectoration into a cup over 3 min. All studies were performed at the same time in the morning. Cell counts were obtained by pipetting the specimen undiluted onto a hemocytometer. A separate aliquot of the undiluted specimen was dried on a slide and stained with Wright stain for differential cell count. The remainder of the specimen was centrifuged at 40,000 × g and the supernatant was used for elastase measurement.

Data Analysis

Parameters for individual patients and group data were expressed as mean values and standard error of the mean or standard deviation as noted. Oropharyngeal colonization was defined as the presence of GNB, Staphylococcus aureus (SA), or yeast. Clearance curves were compared using 95% confidence intervals (CI) to determine if differences were significant. Correlations were assessed by using linear regression, or by Spearman's correlation if data were nonparametric. Multivariate analysis Cox regression analyses were used to investigate the association of clinical risk factors with oral clearance at 60 min. Values were considered significant if p  0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Data

Table 1 lists the sex and clinical characteristics of the patients that may influence salivary flow (medications used) as well as increase risk for colonization and infection (e.g. chronic obstructive pulmonary disease [COPD], nutritional status). The majority of the patients were smokers and had dentures. All other characteristics were present in less than 35% of the patients. The mean age of the patients was 78.9 ± 1.0 (mean ± SEM). The mean Barthel index was 68 ± SEM 2.9 with a maximum of 100, reflecting the high level of physical function of most of these patients. Similarly, the mean MSSE was 21.1 ± SEM 1.1, indicating only mild mental impairment in this group of patients. The high level of both physical and mental function makes this group less likely to have significant serious swallowing problems. None of them were receiving supplemental enteral nutrition.

Microbiologic Data

Table 2 lists the organisms isolated from the buccal isolates. Thirty-seven patients had normal flora (NF). Nine patients had only gram-negative isolates and 17 patients had only yeast. The remainder of the patients' cultures had gram-negative organisms, SA, or yeast present.

Oral Clearance and Colonization

Oral clearance of radiolabel over time for colonized patients and noncolonized patients is shown in Figure 2. The upper line represents colonized patients and shows a significant delay in clearance of radiolabel by 95% CI. The relationship between clearance and colonization status was also confirmed when clearance was compared at different time points in colonized and noncolonized patients. Age, serum albumin, MSSE, and Barthel index were not significantly different in colonized patients compared with noncolonized patients. Retention values at R15 (retention at 15 min), R30 (retention at 30 min), R60 (retention at 60 min), R90 (retention at 90 min), and R120 (retention at 120 min) all indicated significantly different delays in the colonized patients (p < 0.008, p < 0.005, p < 0.013, and p < 0.040, respectively).

View larger version (17K): [in this window] [in a new window]   Figure 2.   Percent retention of the radiolabel over 120 min. Patients with pathogenic colonization had significantly impaired clearance compared to those with NF. The error bars represent 1.96% × SEM (95% CI).

Oral Clearance, Medications, and Clinical Risk Factors for Colonization

In Figure 3 the retention of radiolabel in the oropharynx in patients who were receiving antidepressants is compared with that of those who were not. The percent retention is plotted over time and the upper line represents the patients on medication. Patients medicated with antidepressants had a significant delay in clearance by 95% CI compared with nonmedicated persons. On multivariant analysis the clearance at 60 min was strongly related to being on antidepressants but not to the use of diuretics or antiparkinsonian medications.

View larger version (15K): [in this window] [in a new window]   Figure 3.   Oropharyngeal clearance over time is shown for patients on and off antidepressants. Retention of radiolabel was significantly different in the two groups. The error bars represent 1.96% × SEM (95% CI).

The relationships between functional status, clinical diagnosis, and oral clearance are shown in Figure 4. There was a trend of increased radiolabel retention in patients with COPD, decreased albumin levels, decreased Barthel or MMSE scores, or abnormal dentition (partial or total dentures). Patients with congestive heart failure (CHF) had a trend toward more rapid clearance.

View larger version (44K): [in this window] [in a new window]   Figure 4.   Oropharyngeal clearance over time is shown for patients with COPD, CHF, low albumin, high or low Barthel or MMSE scores, and normal or abnormal dentition (partial or full dentures). The error bars represent 1.96% × SEM (95% CI).

Oral clearance of 11 patients was examined serially 12 mo apart to assess whether any change from increasing age could be detected. These data are shown in Figure 5. There was a decrease in clearance at every time point in the second study; however, these studies were not significantly different by 95% CI.

View larger version (16K): [in this window] [in a new window]   Figure 5.   Oropharyngeal clearance over time is shown for patients (n = 11) at entry into the study and 12 mo later. The error bars represent 1.96% × SEM (95% CI).

Elastase concentrations were significantly elevated in patients with pure GNB compared with patients with NF (mean ± SEM, 10.6 ± 5.7 nM versus 2.2 ± 1.2 nM, p < 0.036). Elastase concentrations were not correlated to oropharyngeal clearance or cell populations.

Salivary Cell Populations: Relationship to Clearance and Colonization

The effect of colonization status on the salivary cell populations is shown in Table 3. The patients with pure gram-negative isolates had a significantly greater percentage of macrophages compared with those with normal flora (p < 0.0016) as did those with mixed pathogens (p < 0.048).

The absolute number of lymphocytes was increased in colonized patients compared with patients with NF (p < 0.048). The number of buccal epithelial cells was also markedly elevated in colonized patients in comparison to those with NF (p < 0.003). Patients with pure gram-negative colonization had the most epithelial cells/ml (p < 0.0006).

Salivary Production and Oral Clearance

Patients who could not produce saliva because of xerostomia were compared with those who could expectorate. In Figure 6 patients with no saliva are represented by the upper line and have significant differences in retention of radiolabel by SEM at each time point compared with those who could expectorate.

View larger version (14K): [in this window] [in a new window]   Figure 6.   Oropharyngeal clearance over time is shown for patients with severe xerostomia (no measurable saliva) and those with nonsevere xerostomia. Patients with severe xerostomia had significantly impaired clearance compared with those without severe xerostomia. The error bars represent SEM.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study demonstrates that elderly patients with pathogenic colonization have significantly impaired clearance of radiolabel compared with those with NF. Age, functional status, and albumin concentrationsfactors thought to be important risk factors for colonizationwere similar in patients with NF and those colonized with pathogens. These findings are consistent with the hypothesis that abnormal clearance may represent the first step of the colonization process.

Earlier investigations are also consistent with the hypothesis that salivary flow and physical clearance may be important elements maintaining normal oral flora. LaForce and coworkers demonstrated that the key event in initial removal of GNB in normal hosts was physical clearance rather than cellular or proteolytic bactericidal events. In their investigation (1), GNB were radiolabeled and sprayed in the oropharynx of normal volunteers. Ninety percent of clearance was shown to be the result of physical removal of bacteria. In a second investigation, they further demonstrated that atropine, an agent which decreases salivary flow, markedly decreased oropharyngeal clearance of radiolabel (2).

These data, which reveal the importance of physical clearance in this experimental model in normal subjects, have important implications for colonization in the elderly. A marked increase in oropharyngeal gram-negative colonization is seen in patients with well-documented decreased salivary flow such as those with Sjögren syndrome or radiation-induced sialadenitis (12). Colonization with yeast is associated with medications that induce xerostomia (13, 15). Thus, it is logical to propose that decreased salivation and oral clearance may be the initiating events for both bacterial and fungal colonization in the elderly.

The decrease in clearance in colonized patients demonstrated in this investigation may be multifactorial in origin. The role of the aging process in decreased salivary flow is controversial but key factors affecting salivation have been identified. These include underlying disease and medications (16- 18). Medications may be particularly important. In a large cross-sectional study over 47% of elderly patients are receiving drugs that can decrease salivary flow (19) These include antidepressants, antiparkinsonian medications, diuretics, antihypertensives, and antihistamines (16). In our investigation clearance was markedly reduced in those patients on antidepressants. Diuretics and antiparkinsonian medications had no significant impact on retention of radiolabel. This is surprising particularly in view of the well-known anticholinergic side effects of the antiparkinsonian drugs. It is possible that both depression and the antidepressants have an additive effect on decreasing salivary flow.

The impact of aging (measured at 1 yr) on oral clearance was assessed by examining patients serially at baseline and 12 mo later. Clearance was delayed; however, this change was not statistically significant. The small number of patients, the short interval between studies, and the wide array of medications and underlying disease processes make causal relationship between age and clearance difficult to isolate. Multivariant analysis revealed no significant relationship between age and clearance. These data support the hypothesis that aging alone may have minimal or no effect on salivary flow.

Salivary elastase activity was increased fivefold in the patients colonized with GNB in this investigation. This finding is similar to that found in postoperative patients who became colonized with gram-negative organisms (12). In this earlier study of ambulatory adult patients, hospitalized for elective cardiac surgery, patients who became colonized had increased elastase concentrations compared with those who did not. Importantly, elastase increased before the colonization. This finding is significant because alterations in epithelial cell bacterial receptors are thought to be promoted by the effect of elastase on cell surface fibronectin (20, 21). These human studies have shown increased bacterial adherence to buccal cells and increased elastase in patients who are at risk for oropharyngeal or tracheobronchial colonization, including postoperative patients or critically ill patients requiring intensive care (20, 21).

Elastase may play a similar role in the colonization process in elderly debilitated patients. We speculate that a decrease in oral clearance may precede increased elastase activity and colonization in both the elderly and postoperative patients. For example, in a dry mouth, bacterial residence simply increases and triggers cellular activity. Elevated elastase concentrations observed in this investigation are a measure of this. Absence of normal clearance of pathogenic organisms may be the seminal event which then initiates a cascade of inflammatory responses leading to colonization.

In addition to the changes in elastase concentrations observed in this study, inflammatory cell populations also changed with colonization. Those patients with respiratory pathogens had a significantly higher percentage of macrophages in the saliva (mean ± SEM, 31 ± 9.0 versus 7.1 ± 1.2, p < 0.003). These macrophages may be salivary in origin or arise from the lower inflamed airway. The absolute number of lymphocytes was increased in patients with pathogens present. The number of buccal epithelial cells was also increased in all colonized patients compared with those with normal flora (p < 0.005). This increase in epithelial cells most likely represents increased oropharyngeal inflammation and resultant desquamation of the epithelial surface.

In summary, we have demonstrated significant impairment in oral clearance in elderly patients colonized with pathogenic organisms. We have further shown that patients colonized with GNB had increased elastase activity and buccal desquamation compared with patients with NF. These data may have important implications for future studies in patient populations at risk for lower respiratory tract infection. Alteration in oral clearance in critically ill patients and perioperative patients, as well as in the elderly, may be the primary event leading to bacterial growth, influx of inflammatory cells, mediator-induced epithelial injury, and ultimately, pneumonia. These findings support prior work that suggests oral health plays a role in colonization (22). Longitudinal studies with quantitative cultures in populations at risk are needed to determine the temporal relationship of decreased clearance to the colonization process to determine if it is causal or follows inflammation. Investigations of therapeutic intervention aimed at improving oral clearance will provide further insight into the mechanisms of oropharyngeal colonization and pneumonia. Such interventions may hold significant promise in decreasing the incidence of nosocomial pneumonia, with powerful benefits with regard to cost, morbidity, and mortality.