Diagnosing Dyspneic Older Adult Emergency Department Patients Pilot Study: Diagnoses and Potential Role of Antimicrobial Peptides

Abstract

Study Objectives:

Pneumonia, chronic obstructive pulmonary disease (COPD), and heart failure (HF) exacerbations can present similarly in the older adult in the Emergency Department (ED), leading to sub-optimal treatment from over- and under-diagnosis. There may be a role for antimicrobial peptides (AMPs) in improving the accurate diagnosis of pneumonia in these patients.

Methods:

This pilot was a prospective, observational cohort study of older adults (aged ≥65 years of age) who presented to the ED with dyspnea or elevated respiratory rate. To identify biomarkers of pneumonia, serum levels of white blood cell count, procalcitonin (PCT), and antimicrobial peptides (human beta defensin 1 and 2 [HBD-1, -2], human neutrophil peptides 1–3 [HNP1–3] and cathelididin [LL-37]) were compared between those with and without pneumonia. Criterion standard reviewers retrospectively determined the diagnoses present in the ED.

Results:

Three hundred ninety-one patients were screened, 140 were eligible, and 79 were enrolled. Based on criterion standard review, pneumonia was present in 10 (12.7%), COPD in 9 (11.4%) and HF in 31 (39.2%) with a co-diagnosis rate of 10.1% by criterion standard review. Comparatively, emergency medicine attending physicians diagnosed pneumonia in 16 (20.3%), COPD in 12 (15.2%), and HF in 30 (38.0%) with co-diagnosis rate of 15.2%. Emergency physicians agreed with criterion standard diagnoses in 90% of pneumonia, 75% of COPD and 65% of HF diagnoses. Differences in leukocyte count (p<0.01) and two novel AMPs (DEFA5 (p=0.08) and DEFB2 (p=0.09)) showed promise for diagnosing pneumonia.

Conclusions:

Emergency physicians continue to have poor diagnostic accuracy in dyspneic older adult patients. Serum AMP levels are one potential tool to improve diagnostic accuracy and outcomes for this important population and require further study.

INTRODUCTION

Dyspnea is the second leading cause of US emergency department (ED) visits in older adult patients aged ≥65 years.¹ In this population, dyspnea itself is an independent predictor of both morbidity and mortality² and presents diagnostic challenges to emergency physicians.^3–7 Diagnostic inaccuracy in dyspnea disproportionately affects older adults.⁸ Approximately 1 in 5 dyspneic older adults have a diagnosis missed in the ED³ and 10–31% are treated for ≥1 of^6,8 the three most common causes of dyspnea (pneumonia, acute exacerbation of chronic obstructive pulmonary disease [COPD], and acute exacerbation of heart failure [HF]).³ In over half of older dyspneic ED patients admitted with one of these three diagnoses, the ED and hospital diagnosis do not agree.⁸

Accurately diagnosing pneumonia in older adults is particularly challenging to emergency physicians due to atypical presentations,^9,10 decreased sensation of dyspnea,¹¹ inaccuracy of chest radiographs (60% sensitivity, 76% specificity),¹² comorbidities which are potential etiologies of symptoms (e.g., CHF or COPD),¹³ overlap of symptoms between pneumonia and other conditions, failure of clinical prediction rules for other pulmonary disease (e.g. pulmonary embolism),^14–16 and inadequate accuracy of available biomarkers (e.g. procalcitonin).¹⁷ Emergency physician diagnostic inaccuracy in these patients has been associated with increased admission rate, longer hospital length of stay, and death or re-hospitalization within one year.⁶ Less than 20% of ED physicians report that they are very confident in diagnosing pneumonia in older adults⁷ and they are less likely to follow pneumonia diagnostic criteria in these patients.¹⁸ As a result, co-diagnosis (e.g combined diagnosis of pneumonia, HF and/or COPD) occurs in 25% of patients admitted with an ED diagnosis of pneumonia.⁵ and 32% of admitted patients diagnosed with pneumonia are simultaneously treated for CHF or COPD.⁴ As many as one in five (22%) patients may have delayed administration of antibiotics.¹⁹ Conversely, 37% of patients receiving antibiotics in the ED for pneumonia do not have a chest x-ray consistent with pneumonia,²⁰ and up to 73% of ED nursing home patients given antibiotics for pneumonia do not meet current diagnostic criteria.²¹

Biomarkers are a potential tool to improve pneumonia diagnosis. Antimicrobial peptides (AMPs) have shown promise in the ED older adult population for diagnosing other infections.²² AMPs are proteins that are a key part of the innate immune system which responds within minutes to a microbial assault by bacteria, viruses and fungi.²³ AMPs have been studied in bronchoalveolar lavage (BAL) samples^24,25 and demonstrated promise in the diagnosis of patients with dyspnea, but data on more readily available biological fluids such as serum levels is more practical because BAL is not available in the ED.

Because pneumonia, COPD and HF are common,^3,4,8 frequently misdiagnosed^4,5,8 and associated with high morbidity/mortality⁶ in older adult ED patients, this prospective pilot study sought to identify the degree of over- and under-diagnosis of pneumonia, COPD, HF in a dyspneic older adult ED population. The primary goal of this study was to prospectively describe diagnosis patterns a dyspneic older adult population. We also sought to perform an exploratory analysis to determine if serum AMP levels should be studied further to improve pneumonia diagnostic accuracy of elderly ED patients.

METHODS

This pilot was a prospective, observational cohort study of ED patients ≥ 65 years of age who presented with dyspnea. Dyspnea was defined as either a chief complaint consistent with dyspnea or shortness of breath or as two measured respiratory rates (RR) ≥20 breaths per minute (bpm) within 4 hours of presentation. RR was used as the sensation of dyspnea can be decreased in older adults,^11,26 and a cutoff RR of ≥20 bpm is used in the validated Emergency Severity Index (ESI).²⁷ Patients were excluded if they had active cancer, immunosuppression, trauma activation, incarceration or suicidal ideation. Informed consent was obtained from all patients. We allowed for consent by appropriate proxy to prevent selection bias. This study was approved by the Institutional Review Board at the institution of enrollment.

Patients were enrolled by trained research assistants between 7AM and 11PM in a single, US academic ED with more than 80,000 annual visits, 18% of are older adults. Enrolled patients completed a survey in the ED administered by trained research assistants collecting demographics and symptoms. At disposition, the emergency medicine attending physician completed a survey regarding the most likely cause of presentation. Trained research assistants completed a chart review to document presenting vital signs, ED treatment, ED diagnoses, disposition, hospital treatment, hospital diagnoses and diagnostic results. Chart review methods included 1 hour training and use of a codebook.

Blood samples were collected in the ED, centrifuged and then serum was stored at −80° Celsius until analyzed using enzyme-linked immunosorbent assays (ELISA). AMPs included in this pilot study were chosen based on our prior experience in older adult UTIs²² and prior studies of BAL specimens.^24,25 They included HNP 1–3 Human ELISA (Hycult Biotech- Plymouth Meeting, PA; catalog number HK317–01), LL-37 Human ELISA (Hycult Bioteck; catalog number HK317–01), Human Alpha Defensin-5 ELISA (Cloud Clone Corporation- Katy, TX; catalog number SEB912Hu) and Human Beta Defensin 2 ELISA (Cloud Clone Corporation; catalog number SEA72Hu). Samples were run using a Molecular Devises (San Jose, CA) Flex 3 plate reader. Samples with levels below the limit of detection (LOD) were assigned the LOD/$\surd 2$.²⁸ We obtained serum leukocyte and procalcitonin in our facility’s clinical laboratory.

Presence or absence of the three diagnoses of interest (pneumonia, COPD, HF) were determined by criterion standard reviewers by review of the electronic medical record and patient survey.³ The criterion standard reviewers were board-certified physicians with expertise in geriatrics, emergency medicine, cardiology, heart failure, infectious diseases, and pulmonology. They had access to the entirety of the patient chart including ED visit, inpatient visit (if applicable) and subsequent visits. Two reviewers adjudicated each patient chart. If there was disagreement on the diagnoses of interest, a third reviewer adjudicated. This criterion standard was compared to ED physician impression.

Sample size was calculated for a larger study based focused on the co-diagnosis and co-treatment of the three conditions of interest in which we under-enrolled patients with a diagnosis of pneumonia (anticipated 30 in our sample). As it was not powered to evaluate the differences between AMPs, this was an exploratory analysis. We report n (%) for categorical variables and mean (95% confidence intervals) for continuous variables. T-tests were used to compare laboratory values with log transformation where appropriate. For significant laboratory values, sensitivity and specificity are reported using established normal values or cut-points determined based upon receiver operating curves (AMPs). Agreement between adjudicators was measured by Cohen’s Kappa. Data management was completed in SAS 9.4 (SAS Institute Inc., Cary, NC) and data analyses in STATA 15 (StataCorp, College Station, TX).

RESULTS

Patients were enrolled from October 2017 to June 2018. Three hundred ninety-one patients were screened, 140 were eligible, 81 were enrolled and 79 with blood samples were included in final study analyses. The most common reason eligible patients did not enroll was patient/family refusal (95.0%). The majority of enrolled patients were eligible by chief complaint only (79.8%) and few were eligible by both criteria (7.6%). Eligible patients who did not enroll were similar in age to patients who did enroll (76.6 years (95% confidence interval (CI) 74.3–79.0) vs. 73.0 (71.4–74.5)). Characteristics of enrolled patients are shown in Table 1.

Table 1.

Demographic characteristics of older ED patients with dyspnea (n=79). All data presented as n (%) unless otherwise noted.

Patient characteristic	n (%)
Age, years [mean (std dev)]	73.1 (7.0)
Female	35 (44.3)
Race
White/Caucasian	63 (79.8)
African American	15 (19.0)
Other/unknown	1 (1.3)
Emergency severity index (ESI) score
1 (most urgent)	4 (5.1)
2 (emergent)	70 (88.6)
3 (urgent)	4 (5.1)
4 (less urgent)	0 (0.0)
5 (non-urgent)	0 (0.0)
Unknown	1 (1.3)
ED Disposition
Admit to floor	61 (77.2)
Admit to intermediate or intensive care	6 (7.6)
Discharge home	10 (12.7)
ED observation unit	2 (2.5)
Place of residence
Home/apartment	72 (91.1)
Assisted living	4 (5.1)
Rehabilitation	1 (1.3)
Skilled nursing facility/nursing home	2 (2.5)
Criterion standard reviewer diagnosis1
Heart Failure (HF)	31 (39.2)
Pneumonia	10 (12.7)
Chronic obstructive pulmonary disease (COPD)	9 (11.4)
Atypical / Non-cardiac chest pain	6 (7.6)
Arrhythmia	6 (7.6)
Viral respiratory illness	5 (6.3)
Acute coronary syndrome/cardiac chest pain	4 (5.1)
Dyspnea not otherwise specified	4 (5.1)
Anemia	3 (3.8)
Pulmonary HTN	3 (3.8)
Renal disease	3 (3.8)
Pleural effusions	2 (2.5)
Pericardial disease	2 (2.5)
Other	17 (21.3)
Co-diagnosis of pneumonia, COPD and HF	8 (10.1)

¹Not mutually exclusive, 105 diagnoses in 79 patients

At least one of the three diagnoses of interest was present in 42 (53.2%) patients as determined by criterion standard reviewers. Pneumonia was diagnosed in 10 (12.7%), COPD in 9 (11.4%), and HF in 31 (39.2%). We enrolled a lower proportion of pneumonia cases than expected. Pneumonia diagnosis agreement between criterion reviewers occurred in 96.2% (kappa [κ]=0.85); for COPD agreement was 91.1% (κ=0.54); and for HF agreement was 88.6% (κ=0.76). Co-diagnosis was present in 10.1%. (Table 1)

Considering all diagnoses for dyspnea determined by the criterion standard reviewers, HF, pneumonia and COPD remained the most common (Table 1). The emergency medicine attending physician diagnosed pneumonia in 16 (20.3%), COPD in 12 (15.2%), and HF in 30 (38.0%) with a co-diagnosis rate of 15.2%. Emergency physicians had higher rates of under-diagnosis than over-diagnosis of pneumonia, COPD and HF compared to criterion standard reviewers. (Table 2)

Table 2.

ED attending physician diagnosis agreement with criterion review of pneumonia, chronic obstructive pulmonary disease (COPD) and/or heart failure (HF) in 79 dyspneic older ED patients. Data are presented as n (%).

	Criterion standard diagnosis present			Criterion standard diagnosis not present
	ED MD agrees	ED MD disagrees (under-diagnosis)	Total present	ED MD agrees	ED MD disagrees (over-diagnosis)	Total not present
Pneumonia	9 (90)	1 (10)	10 (13)	62 (90)	7 (10)	69 (87)
COPD	7 (78)	2 (22)	9 (11)	65 (93)	5 (6)	70 (89)
HF	20 (65)	11 (35)	31 (39)	38 (79)	10 (21)	48 (61)

We next examined serum makers of infection in patients with pneumonia. Leukocyte count was higher in patients diagnosed with pneumonia compared to those without (p<0.01) but procalcitonin was not significantly different (p=0.42). Among the novel AMPs studied, DEFA5 (p=0.0873) and DEFB2 were higher in patients with pneumonia and trended towards significance (p=0.077). HNP1–3 (p=0.0.247) and LL-37 (p=0.0.608) were not significantly different in patients with pneumonia compared to those without. (Table 3). All samples measured within the ELISA range of detection for HNP1–3, LL37 and DEFA5. One measurement for DEFB2 was below the LOD and was assigned the value of LOD/square root of 2. Seven DEFB2 measurements were higher than the range of detection. Because the standard curve was linear for DEFB2 and the samples were only 1.1–1.7 fold above the detection limit, we extrapolated the values for these samples from the standard curve.

Table 3.

Mean laboratory values in dyspneic older ED patients diagnosed with and without pneumonia. Sensitivity, specificity, negative predictive value (NPV) and positive predictive value (PPV) reported for laboratory studies that are significant or trend towards significance. Data are presented as mean and 95% confidence interval.

	Pneumonia (n=10)		No pneumonia (n=69)		p-value	Sensitivity	Specificity
Leukocyte count (K/uL)1	11.5	(7.0–16.1)	7.8	(7.1–8.4)	<0.01	50.0%	89.9%
Procalcitonin (ng/mL)	0.18	(0.03–0.33)	0.11	(0.05–0.17)	0.42
HNP1–3 (pg/mL)	224.1	(205.1–243.1)	387.6	(210.2–564.9)	0.25
LL-37 (ng/mL)	3.1	(1.8–4.3)	3.1	(2.1–4.0)	0.61
DEFB2 (pg/mL)2	1,072.1	(470.8–1,673.3)	608.9	(432.6–786.2)	0.09	80.0%	44.9%
DEFA5 (pg/mL)3	428.7	(295.3–562.1)	342.0	(316.9–367.1)	0.08	80.0%	44.9%
	Pneumonia (n=10)		No pneumonia (n=69)		p-value
Leukocyte count (K/uL)	11.5	(7.0–16.1)	7.8	(7.1–8.4)	<0.01
Procalcitonin (ng/mL)	0.18	(0.03–0.33)	0.11	(0.05–0.17)	0.42
HNP1–3 (pg/mL)	224.1	(205.1–243.1)	387.6	(210.2–564.9)	0.25
LL-37 (ng/mL)	3.1	(1.8–4.3)	3.1	(2.1–4.0)	0.61
DEFB2 (pg/mL)	1,072.1	(470.8–1,673.3)	608.9	(432.6–786.2)	0.09
DEFA5 (pg/mL)	428.7	(295.3–562.1)	342.0	(316.9–367.1)	0.08
	Sensitivity	Specificity	NPV	PPV
Leukocyte count (K/uL)1	50.0%	89.9%	92.5%	41.7%
DEFB2 (pg/mL)2	80.0%	44.9%	93.9%	17.4%
DEFA5 (pg/mL)3	80.0%	44.9%	91.2%	15.6%

¹Cut point 10.10 for males, 11.19 for females

²Cut point 266.4

³Cut point 313.6

DISCUSSION

This study describes a prospective cohort of dyspneic older adult ED patients, re-demonstrates the diagnostic challenge facing emergency medicine physicians, and identifies two AMPs that have promise for differentiating patients with and without pneumonia in this population. Consistent with previous literature,³ HF, pneumonia and COPD were the most common causes of dyspnea in our population and co-diagnosis was common (10.1%). Compared to prior studies, our study benefits from adjudication of presence of HF, pneumonia and COPD by experts and prospectively identifying patient symptoms rather than relying on chart review.

Emergency physicians regularly care for very ill patients with limited information and time, leading to diagnostic inaccuracy ^3,5 as confirmed in this study. In older adults, this challenge in compounded by atypical presentations,^9,10 decreased sensation of dyspnea,¹¹ presence of multiple comorbidities which are potential etiologies of dyspnea,¹³ failure of clinical prediction rules for pulmonary disease in older adults.^14–16 Recent literature has even questioned the utility of signs and symptoms in this population.²⁹

Therefore, a quantitative test that can be applied quickly and accurately in the older adult population, such as an AMP, or a quantitative test that is part of a decision rule, similar to the use of a d-dimer with the Well’s Criteria for pulmonary embolism, would be well-suited to the ED. Improving diagnostic accuracy in the ED would also be expected to improve the inpatient care and long-term outcomes of these patients.⁶ Our study demonstrates that two AMPs, DEFB2 and DEFA5, may help identify older adult ED patients with pneumonia; the observed trend towards significance is likely due to lower power than planned. Procalcitonin, the only biomarker currently in use for predicting infection in this population, had poor sensitivity in this study. This is the first step in the evaluation of new biomarkers; next steps will include prospective validation and determination of incremental value and clinical utility.³⁰

Although benefiting from a prospective approach which allowed gathering of complete symptom data rather than relying on chart review and used a criterion standard of multiple experts to confirm diagnoses, this study has several limitations. First, these results describe a preliminary cohort of dyspneic older adults in one academic ED, potentially limiting generalizability to other similar, large academic EDs. Additionally, the study was under-powered due to lower enrollment of pneumonia patients than anticipated. We considered enrolling to enrich the population of pneumonia patients but we did not want to bias the study sample. Finally, the specificity of emergency physician diagnosis is artificially elevated in this population as it was clear that some patients’ dyspnea was due to other diagnoses (for example, atrial fibrillation with rapid ventricular response). Future studies should have larger sample size and enroll a cohort of patients suspected of having pneumonia.

CONCLUSION

The prospective preliminary study demonstrates that emergency physicians continue to have poor diagnostic accuracy in dyspneic older adult ED patients when a criterion standard is used to establish the diagnosis. Two novel biomarkers (DEFB2 and DEFA5) are elevated in pneumonia and may ultimately be useful diagnostic markers to help improve diagnostic accuracy and outcomes for this important population and require further study.

IMPACT STATEMENT.

We certify that this work is novel. This work prospectively describes the diagnostic patterns in dyspneic older adults, explores agreement with inpatient physicians and explores the potential role of antimicrobial peptides in the diagnosis of pneumonia.