Hospital admissions for pneumonia more likely with concomitant dental infections

Transcription

1 Clin Oral Invest (2015) 19: DOI /s y ORIGINAL ARTICLE Hospital admissions for pneumonia more likely with concomitant dental infections Brian Laurence & Nee-Kofi Mould Millman & Frank A. Scannapieco & Armin Abron Received: 4 March 2014 /Accepted: 20 October 2014 /Published online: 31 October 2014 # Springer-Verlag Berlin Heidelberg 2014 Abstract Objective The objective of this study is to determine if the presence of dental infection is associated with an increased likelihood of hospital admission following an emergency department (ED) visit among patients diagnosed with pneumonia. We hypothesized that the presence of a dental infection may worsen the clinical symptoms in ED patients diagnosed with pneumonia and are using hospital admission as a marker of worsening clinical severity. Materials and methods We analyzed the data from the 2008 Nationwide Emergency Department Sample and used Poisson regression with robust estimates of variance to obtain prevalence ratios (PRs) with the appropriate adjustments for complex survey sampling. Results In the final multivariable model, there was a 19 % increase in the likelihood of hospital admission following an ED visit among pneumonia patients diagnosed with dental infection compared to those without dental infection (PR= 1.19, 95 % CI= ). In an exploratory multivariable analysis, pneumonia patients diagnosed with dental caries had a 29 % increase in the likelihood of admission compared to those not having dental caries (PR=1.29, 95 % CI=1.23 B. Laurence (*) Department of Restorative Services, Howard University College of Dentistry, Washington, DC, USA brianlaurence2012@gmail.com N.<K. Mould Millman Department of Emergency Medicine, University of Colorado, Anschutz Medical Campus, Aurora, CO, USA F. A. Scannapieco Department of Oral Biology, School of Dental Medicine, University at Buffalo, Buffalo, NY, USA A. Abron Department of Periodontology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA 1.34). These findings remained consistent in a subgroup analysis among patients with less clinically severe forms of pneumonia. Conclusions Dental infections may worsen the clinical symptoms in ED patients with pneumonia increasing their likelihood for hospital admission. Dental caries may be a marker for poor oral hygiene and increased dental plaque rather than serve directly as a source of respiratory pathogens. Clinical relevance The findings suggest that an increased focus on preventive oral health may reduce the need for admission following an ED visit for patients diagnosed with pneumonia. Keywords Dental caries. Dental infection. Emergency treatment. Hospital admission. Pneumonia Published studies report linkages between poor dental health and an increased likelihood of development of pneumonia [1 3]. In a systematic review of the evidence describing the potential association between respiratory diseases and oral health, Azarpazhooh and Leake concluded that there was fair evidence (II-2, grade B recommendation) of an association between poor oral health and the development of pneumonia depending on the oral health indicator used [4]. More specifically, several studies have shown an association between oral plaque, periodontal pathogens, and the development of pneumonia, mostly among hospitalized and nursing home patients [1 3, 5 7]. Scannapieco suggested four possible mechanisms through which oral bacteria could lead to an increase in respiratory pathology [8]. These include (1) the direct aspiration of dental pathogens into the lungs to cause infection, (2) aspiration of periodontal disease-associated enzymes that facilitate the adherence of respiratory pathogens to the airways, (3) hydrolytic enzymes from periodontal pathogens could destroy protective salivary pellicles resulting in fewer non-specific host defense

2 1262 Clin Oral Invest (2015) 19: mechanisms in high-risk subjects, and (4) cytokines from periodontal diseased tissues are aspirated to increase inflammation in the lower airway that may encourage respiratory pathogen colonization in high-risk patients by upregulating the expression of adhesion receptors on mucosal surfaces. In addition, other unknown mechanisms may be at play. Previous analysis of the Nationwide Emergency Department Sample (NEDS) found an association between the presence of a dental infection and an increased likelihood of hospital admission among adult patients with sickle cell disease if the dental infection occurred concurrently with a sickle cell crisis [9]. Using the same robust methodology employed in that study, we sought to determine if the presence of dental infection was associated with an increased likelihood of hospital admission among adult patients diagnosed with pneumonia. Given the growing evidence of associations between poor oral health and an increased likelihood of the development of pneumonia, we hypothesized that the presence of dental infections increases the susceptibility in patients diagnosed with pneumonia. We have used hospital admission as a marker for worsening pneumonia clinical severity. As secondary exploratory aims, we also sought (1) to determine the magnitude of the associations between dental caries and hospital admission as caries may be a proxy for poor dental plaque control and (2) to determine the magnitude of the associations between dental caries and dental infection and hospital admission of patients with less clinically severe forms of pneumonia or categories of pneumonia that would not necessarily warrant hospital admission following an ED visit. Materials and methods We obtained data from the Nationwide Emergency Department Sample (NEDS). The NEDS contains information about emergency department (ED) visits across the United States of America (USA) and includes information on patient, hospital, and clinical characteristics as well as reasons for the ED visit. The NEDS is the largest all-payer ED database that is currently publicly available in the USA. This analysis was performed on data released for the year For the year 2008, the NEDS captured information from over 28 million ED visits at 980 hospitals that approximated a 20 % stratified sample of US hospital-based EDs. Importantly, the NEDS provides weights that are used to allow the calculation of nationallevel estimates. The presence or absence of dental infection and dental caries was our two independent variables of interest. These were defined as two separate dichotomous variables and each independent variable was analyzed separately. Dental infection was defined as the presence of any of the following International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes for periapical/pulpal infections or periodontal infections, to and to We chose not to separate the periodontal from the periapical/pulpal infections as we believed it was highly unlikely that the ED healthcare workers making the diagnosis would be able to make such a distinction with a high level of reliability given the lack of specific oral health training and uncertainty about whether or not the ED facility provided access to periapical or panoramic radiographs. For the exploratory analysis, dental caries was defined as the presence of any of the following International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes: 521.0, , , , , and A detailed list of specific ICD-9-CM codes and their corresponding diseases and injuries can be found at The dependent variable or outcome measure was ED visit disposition that was dichotomized to represent whether or not the ED visit ended in admission versus being treated in the ED and released. All other ED event outcomes were excluded including ED visits in which the patient was transferred to another short-term hospital (1.3 % of ED visits), ED visits in which the patient was not admitted to the same hospital but the final destination was unknown (0.7 % of ED visits), and ED visits in which the patient died in the ED (0.2 % of visits). The ED visits of individuals with pneumonia were identified by using the single level Clinical Classifications Software (CCS) code for pneumonia which excludes pneumonia caused by tuberculosis or sexually transmitted diseases. The CCS collapses diagnosis codes from the ICD-9-CM into 285 mutually exclusive diagnosis categories from over 14,000 individual diagnosis codes. Based on a comprehensive a priori review of the literature, we included in our analysis the clinical variables that could lead to an increased likelihood of hospital admission among adult patients with pneumonia as well as the relevant demographic and hospital characteristics [10, 11]. The demographic characteristics included in the analyses were age, gender, the primary expected insurance payer, and the median household income for the patients zip code. Potentially confounding clinical characteristics were identified using the relevant ICD-9-CM codes and included the presence or absence of the following: an injury (identified using CCS codes 225 through 244), cerebrovascular disease (430, 431,432, ,433, ,434, ,436,437, ), chronic kidney disease or renal failure (585, ), bacteremia/septicemia (790.7, 038.0,038.1,038.2,038.3,038.4, ,038.8,038.9), HIV (042), heart failure ( ), chronic liver disease or cirrhosis ( ), and malignant neoplasm of the lung, trachea, or bronchus (162, ). The hospital characteristics that were included as potential confounders were the hospital s trauma center level (I, II, or III) and the hospital s teaching status (teaching or non-teaching). We used multivariable Poisson regression with robust estimates of variance to estimate the prevalence ratios (PRs) and

3 Clin Oral Invest (2015) 19: % confidence intervals (CIs) for the association between dental infection and ED visit discharge disposition (admitted vs. released) and then repeated the analysis with dental caries as the independent variable instead of dental infection for the secondary analysis. Although logistic regression is a common regression analysis technique used when analyzing binary outcome data, the use of a Poisson regression with robust estimates of variance provides more accurate estimates when the outcome has prevalence greater than 10 % [12, 13]. It, therefore, avoids the overestimation of risks that can be produced by logistic regression when analyzing common dichotomous outcomes. We adjusted for the effects of clustering, stratification, and weighting as appropriate for the analysis of complex survey data and used a Taylor series linearization to adjust for the standard errors. We also performed an analysis in which we used a combination of respiratory failure and/or respiratory distress to create a stratum of high versus low clinical severity for pneumonia. We used the presence of any of the following ICD-9- CM codes to create a stratum of high clinical severity: acute respiratory failure (518.81), acute respiratory distress (518.82), chronic respiratory failure (518.83), pulmonary insufficiency following trauma and surgery (518.5), and acute and chronic respiratory failure (518.84) [14, 15]. We did not perform an analysis among patients with high clinical severity as they all would have had the outcome, hospital admission, and the analysis was therefore only performed among patients that did not have any of the codes mentioned for respiratory distress or respiratory failure. The ICD-9-CM codes did not allow us to identify and thus stratify analyses according to whether the pneumonia was community-acquired pneumonia (CAP) or nosocomial pneumonia which includes hospital-acquired pneumonia (HAP) or ventilator-acquired pneumonia (VAP) even though these categories have no association with pneumonia severity. As we had already adjusted for variables that were associated with pneumonia severity or hospital admission for patients with pneumonia, this secondary analysis was performed to supplement an already robust analysis and to examine the strength of the association among those with a lower clinical severity. Each individual ED visit served as the unit of analysis and only adults 18 years old and over were included in the analyses. We excluded from the analysis those ED visits with missing values for any of the independent, dependent, or confounding variables. A p value of 0.05 was considered statistically significant, and all analyses were carried out using STATA software, v. 11 (StataCorp, College Station, Texas). Results Table 1 shows the selected characteristics of the sample for pneumonia patients with and without dental infection. Application of the appropriate sample weights for analyses led to an estimated 2,560,752 visits for patients without dental infection and an estimated 637 visits for patients with dental infection. The presence of dental infection in pneumonia patients was associated with a greater likelihood of admission following an ED visit, 87.5 versus 76.7 % in pneumonia patients without dental infection. There were significant differences in the mean age, gender distribution, and primary expected payer for those patients with and without dental infection. Patients with dental infection were younger and more likely to be male, and pneumonia patients with dental infection were less likely to have Medicaid as the primary expected payer. The clinical characteristics were significantly different for three of the included clinical variables with pneumonia patients with dental infection having a higher prevalence of bacteremia and HIV and a lower prevalence of heart failure. For the hospital characteristics, patients with dental infection were more likely to be seen at a metropolitan teaching hospital as compared to a metropolitan non-teaching hospital. Table 2 shows the selected characteristics for pneumonia patients with and without dental caries. The distribution of clinical, demographic, and hospital characteristics was similar to those patients with and without dental infection except that significant differences were observed in the distribution of median household income and pneumonia patients with dental caries had a lower prevalence of chronic kidney disease and heart failure but a higher prevalence of HIV as compared to pneumonia patients without dental caries. The adjusted prevalence ratios for the relationship between dental caries and hospital admission and dental infection and hospital admission are shown in Table 3. In the final model, patients with dental caries were 29 % more likely to be admitted compared to patients without dental caries and patients with dental infection were 19 % more likely to be admitted compared to patients without dental infection. Table 4 shows the analysis among patients with less clinically severe forms of pneumonia (i.e., those without respiratory distress or respiratory failure). Pneumonia patients with dental caries were 35 % more likely to be admitted while the estimated increase in the likelihood of admission for pneumonia patients with dental infection showed no change. Discussion This study provides evidence of an association between dental infection and increased likelihood of hospital admission among adult patients diagnosed with pneumonia visiting an emergency department in the USA. Applying robust analysis to the data and controlling for pertinent demographic, clinical, and healthcare institution factors, the study confirmed our primary hypothesis and demonstrated a positive association

4 1264 Clin Oral Invest (2015) 19: Table 1 Pneumonia study characteristics according to the presence or absence of dental infection for adults 18 years old and over based on national level estimates obtained using weighted data *Standard deviations shown next to mean estimates Pneumonia with dental infection (n=637) Pneumonia with no dental infection (n=2,560,752) p value Mean (SD) age in years* 55.6 (1.50) 65.0 (0.03) <0.001 Admitted following ED visit (%) <0.01 Gender (%) Male <0.05 Primary expected payer (%) Medicaid <0.001 Medicare Private Self-pay No charge Other Median household income (%) $1 38, $39,000 47, $48,000 62, $63,000 or more Clinical characteristics Indicator for injury code (%) Present Cerebrovascular disease (%) Present Chronic kidney disease or renal failure (%) Present Bacteremia or septicemia (%) Present HIV (%) Present <0.001 Heart Failure (%) Present <0.05 Chronic liver disease or cirrhosis (%) Present Malignant neoplasm of lung, trachea, or bronchus (%) Present Hospital characteristics Trauma level (%) Non-trauma center Trauma level I Trauma level II Trauma level III Trauma level I or II, collapsed category Trauma level I, II, or III collapsed category Hospital teaching status (%) Metropolitan non-teaching <0.05 Metropolitan teaching Non-metropolitan

5 Clin Oral Invest (2015) 19: Table 2 Pneumonia study characteristics according to the presence or absence of dental caries for adults 18 years old and over based on national level estimates obtained using weighted data *Standard deviations shown next to mean estimates Pneumonia Pneumonia with p value with dental no dental caries (n=1367) caries (n=2,560,023) Mean (SD) age in years* 53.4 (1.1) 65.0 (0.03) <0.001 Admitted following ED visit (%) <0.01 Gender (%) Male <0.001 Primary expected payer (%) Medicaid <0.001 Medicare Private Self-pay No charge Other Median household income (%) $1 38, <0.001 $39,000 47, $48,000 62, $63,000 or more Indicator for injury code (%) Present Cerebrovascular disease (%) Present Chronic kidney disease or renal failure (%) Present <0.01 Bacteremia or septicemia (%) Present HIV (%) Present <0.001 Heart Failure (%) Present <0.001 Chronic liver disease or cirrhosis (%) Present Malignant neoplasm of lung, trachea, or bronchus (%) Present Hospital characteristics Trauma level (%) Non-trauma center <0.001 Trauma level I Trauma level II Trauma level III Trauma level I or II, collapsed category Trauma level I, II, or III collapsed category Hospital teaching status (%) Metropolitan non-teaching <0.01 Metropolitan teaching Non-metropolitan between dental infection as well as dental caries with hospital admission. These findings still held true in a subset analysis among patients with less clinically severe forms of pneumonia. From our extensive review of the literature, our study is the first we know of to examine the relationship between dental infections and the likelihood of hospital

6 1266 Clin Oral Invest (2015) 19: Table 3 Adjusted prevalence ratios for the relationship between having dental caries or a dental infection and hospital admission among adults with pneumonia Variable Dental caries Dental infection PR 95 % CI PR 95 % CI Dental outcome Present (ref. Absent) 1.29* * *p< This table shows the result from the multivariable model that adjusts for age, gender, the primary expected payer, the median household income, and the presence or absence of the following clinical variables: injury, pulmonary hypertension, chronic kidney disease or renal failure, bacteremia or septicemia, HIV, heart failure, chronic liver disease or cirrhosis, malignant neoplasm of lung, trachea or bronchus, and the following two hospital characteristics: teaching status of hospital and trauma center level admission among adult patients with pneumonia. Our findings support the hypothesis that the presence of dental infections may serve as a risk indicator or modifier in patients diagnosed with pneumonia resulting in hospital admission, shown to be a valid marker of worsening clinical severity in patients with pneumonia [10, 16]. We were unable to distinguish the classification of pneumonia based on the available ICD-9-CM codes in the NEDS database. Hospital acquired (HAP) or nosocomial pneumonia is defined by a pneumonia that occurs at least 48 h after the patient has been admitted. Other classifications include healthcare-associated pneumonia (HCAP), which refers to pneumonia associated with healthcare risk factors such as residing in a nursing home, community-acquired pneumonia (CAP), and ventilator-associated pneumonia (VAP). The median hospital risk-standardized (all cause) pneumonia readmission rates among Medicare patients in the USA were reported as 18.2 % (95 % CI= %) in 2008 [17]. A separate, large multicenter trial of older patients with community-acquired pneumonia (CAP) in the USA found a Table 4 Adjusted prevalence ratios for the relationship between having dental caries or a dental infection and hospital admission among adults with pneumonia characterized as being of low clinical severity Variable Dental caries Dental infection PR 95 % CI PR 95 % CI Dental outcome Present (ref. Absent) 1.35* * *p< This table shows the result from the multivariable model that adjusts for age, gender, the primary expected payer, the median household income, and the presence or absence of the following clinical variables: injury, pulmonary hypertension, chronic kidney disease or renal failure, bacteremia or septicemia, HIV, heart failure, chronic liver disease or cirrhosis, malignant neoplasm of lung, trachea or bronchus, and the following two hospital characteristics: teaching status of hospital and trauma center level 12 % 30-day readmission rate [18]. We were unable to locate any published US studies that discuss the relative prevalence or severity of community acquired to nosocomial pneumonia in emergency departments. Similarly, no data were available for hospital readmission rates classified by the type of pneumonia. We are therefore left to conclude that potentially % of our included study populations were return visits for pneumonia and since we controlled for clinical severity and clinical comorbidities, the effect of this potentially confounding factor was minimized in our analysis. There is evidence that the oral microflora together with risk factors for pneumonia provide a biologically plausible pathway for oral biofilms in the development of pneumonia [19, 20]. The studies to date on the associations between oral pathogens and pneumonia have been largely completed on patients with nosocomial pneumonia. However, one study on a community-dwelling population on the relationship between periodontal status and acute respiratory disease found no association while another study found that the presence of periodontal infections increased the risk of developing community-acquired pneumonia [21, 22]. Prospective studies have also shown that increases in plaque scores are associated with the development of ventilatorassociated pneumonia (VAP) or respiratory pathology [3, 5, 6, 23, 24]. Scannapieco suggests that poor dental hygiene and periodontal diseases may support oropharyngeal colonization by respiratory pathogens such as Escherichia coli and Klebsiella [25]. Dental plaque may therefore serve as a reservoir for the pathogens that could lead to the development of pneumonia, through aspiration or hematogenous spread. In summary, most of the evidence to date suggests that some of the bacteria associated with oral plaque and periodontal disease are risk factors for the development of nosocomial pneumonia [26]. As there is a little evidence to suggest that dental caries, or the microflora associated with dental caries, is directly related to the development of pneumonia, the results of our analysis suggest that dental caries may be a marker for poor oral hygiene or oral care. Thus, potential respiratory pathogenic bacteria that are increased in dental plaque rather than specific cariogenic bacteria serve directly as a source of respiratory pathogens. Unfortunately, we did not have available any measures of oral hygiene or hygiene practices to explore whether caries is a surrogate for poor oral hygiene habits that could lead directly to increased prevalence of dental plaque and an increase of the bacteria shown to be more likely associated with the development of pneumonia. Strengths of this study include the large sample size that permitted the calculation of estimates for the prevalence ratios with small confidence intervals that were observed in the final multivariable models and the robust methodology used in the analyses. In addition, the NEDS data by nature of its nationally representative design allows for a study with increased external validity. As always, however, there are limitations to

7 Clin Oral Invest (2015) 19: any observational study. First and foremost, the reliability and validity of those making the diagnosis of dental caries and dental infection is a concern. Unless it is the major presenting complaint, ED physicians most likely will not perform a dental exam nor document an incidental finding of dental caries. Infections like dental abscesses, especially those causing cellulitis, are obviously more concerning and stand a greater chance of getting appropriately documented and coded as a diagnosis. We could find no studies discussing the reliability or validity of ED physicians in diagnosing dental caries or dental infections and we can only conclude that there is likely misclassification occurring among both the dental caries and dental infection diagnoses. We should, therefore, add that there is a possibility that our findings are spurious and driven by high levels of misclassification of dental caries and dental infection. There were also some important differences between our comparison groups that deserve some discussion, and even though we adjusted for the relevant variables in our analysis, there always remains the possibility of residual confounding. In comparing admitted pneumonia patients with dental infection and those without infections, we observed that the mean age was lower and that there was a predominance of males in the dental infection group. With regards to clinical characteristics, the prevalence of a positive HIV status was higher in the dental infection group, not surprising given HIV patients are more likely to be immunosuppressed thereby disposing them to more frequent infections. The prevalence of heart failure was significantly lower in the dental infection group, and this was an expected finding given the lower mean age of this group. Interestingly, the prevalence of bacteremia and septicemia was significantly higher in pneumonia patients with dental infections. While we cannot draw causal relationships from our analysis, we postulate that the presence of potential respiratory pathogens in the oral cavity (through hematogenous spread or via aspiration) may be causing this increased rate of bacteremia or septicemia, thereby worsening clinical severity and increasing the likelihood of admission rates in this group. Similarly, in comparing those pneumonia patients with dental caries to those without dental caries, we observed a lower prevalence of kidney failure in those with dental caries and this was not unexpected given the lower mean age of that group. Finally, we observed that patients with dental caries tended to be poorer and less likely to have Medicaid than those without dental caries, but we could find no studies showing an association between income level and Medicaid status with the likelihood of admission for patients with pneumonia so we are not able to make any statements about the effect of low income and a higher prevalence of Medicaid in the likelihood of admission. As stated before, dental caries may be a marker for poor oral hygiene and increased levels of dental plaque or it may be that both dental caries and dental infection are associated with some unmeasured lifestyle factors which we did not account for in our analyses that may be associated with an increased likelihood of hospital admission for patients with pneumonia. However, to account for this, we did further analyses including more comprehensive models than those shown in which we adjusted for more possible risk factors, those we considered to be risk factors for admission but weaker risk factors than the previously included variables, and we observed little to no change in the estimated prevalence ratios. We also performed another analysis where we combined dental infections and dental caries into one variable in an attempt to control to some extent for the potential misclassification between the two variables. Following this, we observed that pneumonia patients with a diagnosis of dental infection or dental caries were 27 % more likely to be admitted compared to those without a diagnosis of dental infection or dental caries (PR=1.27, 95 % CI= ). In addition, we could not differentiate among the various classifications of pneumonia using the available ICD- 9-CM codes and could therefore not perform stratified analysis to determine the influence of dental infection on pneumonia admission rates. Finally, in our previous study on the association between dental infection and hospital admission in patients with sickle cell disease, we provided estimates of costs savings for the prevention of dental infection. However, we are not willing to provide such estimates for this study given the wide variation in hospital costs between patients with CAP, HAP, or VAP and our inability to differentiate between them using the available ICD-9-CM codes [27, 28]. Conclusion This study is one of the first to show a positive relationship between the presence of a dental infection and the likelihood of hospital admission among adult patients with pneumonia visiting emergency departments in the USA. As with our earlier study on the association between dental infection and hospital admission among adult patients with sickle cell disease, the findings suggest that an increased focus on preventive oral health or on the early identification and treatment of dental infections may reduce the need for admission following ED visits for patients with specific diseases. We recommend that future prospective studies be designed so that the presence of dental plaque and dental infection can be measured with a high degree of validity and reliability to determine if dental plaque is an independent risk factor for the development of pneumonia in patients visiting emergency departments in the USA. Acknowledgments We would like to acknowledge the support of Ralph Katz, BS, DMD, MPH, PhD. Professor, Epidemiology and Health Promotion, New York University College of Dentistry. New York, NY. USA. Dr. Katz kindly agreed to review a version of this manuscript before it was submitted. Conflict of interest interest The authors declare that they have no conflict of

8 1268 Clin Oral Invest (2015) 19: References 1. Terpenning M et al (2001) Aspiration pneumonia: Dental and oral risk factors in an older veteran population. J Am Geriatr Soc 49(5): Langmore S et al (1998) Predictors of aspiration pneumonia: How important is dysphagia? Dysphagia 13(2): Shi Z et al (2013) Oral hygiene care for critically ill patients to prevent ventilator-associated pneumonia. Cochrane Database Syst Rev 13(8):CD Azarpazhooh A, Leake J (2006) Systematic review of the association between respiratory diseases and oral health. J Periodontol 77(9): Munro C et al (2006) Oral health status and development of ventilator-associated pneumonia: a descriptive study. Am J Crit Care 15(5): Eom J et al (2014) The impact of a ventilator bundle on preventing ventilator-associated pneumonia: a multicenter study. Am J Infect Control 42(1): Tada A, Miura H (2012) Prevention of aspiration pneumonia (AP) with oral care. Arch Gerontol Geriatr 55(1): Scannapieco F (1999) Role of oral bacteria in respiratory infection. J Periodontol 70(7): Laurence B, Haywood C Jr, Lanzkron S (2013) Dental infection increase the likelihood of hospital admissions among adult patients with sickle cell disease. Community Dent Health 30(3): Fine M et al (1997) A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 336(4): Chew K et al (2011) The PORT score may be a valid tool to predict mortality and need for hospital admission in HIV-infected patients with CAP. AIDS Patient Care STDS 25(5): Barros A, Hirakata V (2003) Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 20(3): Cummings P (2009) The relative merits of risk ratios and odds ratios. Arch Pediatr Adolsc Med 163(5): Phua J et al (2009) Validation and clinical implications of the IDSA/ ATS minor criteria for severe community-acquired pneumonia. Thorax 64(7): Espana P et al (2006) Development and validation of a clinical prediction rule for severe community-acquired pneumonia. Am J Respir Crit Care Med 174: Renaud B et al (2007) Routine use of the Pneumonia Severity Index for guiding the site-of-treatment decision of patients with pneumonia in the Emergency Department: a multicenter, prospective, observational, controlled cohort study. Clin Infect Dis 44: CMMS (2011) Centers for Medicare and Medicaid Services. Medicare Hospital Quality Chartbook Performance report on Readmission Measure for Acute Myocardial Infarction, Heart Failure and Pneumona. Sept 15, Jasti H et al (2008) Causes and risk factors for rehospitalization of patients hospitalized with community-acquired pneumonia. Clin Infect Dis 46: Sarin J et al (2008) Reducing the risk of aspiration pneumonia among elderly patients in long-term care facilities through oral health interventions. J Am Med Dir Assoc 9(2): Shay K (2002) Infectious complications of dental and periodontal diseases in the elderly population. Clin Infect Dis 34(9): Scannapieco F, Papandonatos G, Dunford R (1998) Associations between oral conditions and respiratory disease in a national sample survey population. Ann Periodontol 3(1): de Melo Neto J et al (2013) Periodontal infections and communityacquired pneumonia: a case-control study. Eur J Clin Microbiol Infect Dis 32(1): Scannapieco F, Stewart E, Mylotte J (1992) Colonization of dental plaque by respiratory pathogens in medical intensive care patients. Crit Care Med 20(6): Fourrier F et al (1998) Colonization of dental plaque: a source of nosocomial infections in intensive care unit patients. Crit Care Med 26(2): Scannapieco F, Mylotte J (1996) Relationships between periodontal disease and bacterial pneumonia. J Periodontol 67(10 Suppl): Paju S, Scannapieco F (2007) Oral biofilms, periodontitis, and pulmonary infections. Oral Dis 13(6): File TJ, Marrie T (2010) Burden of community-acquired pneumonia in North American adults. Postgrad Med 122: Restrepo M et al (2010) Economic burden of ventilator-associated pneumonia based on total resource utilization. Infect Control Hosp Epidemiol 31: