a b s t r a c t

Background: A patient’s vital signs are all inextricably interrelated, and together provide critical informa- tion regarding hemodynamic and physiological status. Yet, the precise relationship between body tem- perature (T) and heart rate (HR) in adults remains a fundamental gap in our knowledge.

Methods: We performed a retrospective secondary analysis of (1) electronic medical records from a large academic center (annual ED census of 110,000) and (2) the National Hospital Ambulatory Medical Care Survey , a large CDC-sponsored weighted sample of U.S. EDs and our own large tertiary care ED, extracting Demographic and clinical data including vital signs.

Results: We included 8715 local ED visits and approximately 123.3 million estimated national adult ED visits. Mean T was 36.9 ?C, and 5.2% of patients had a T over 38 ?C. Mean (SD) HR was 93.3 bpm, 28% had a HR over 100 bpm. Males had significantly lower HR than females (coefficient -1.6, 95%CI -2.4 to -0.8), while age was negatively associated with HR (coefficient -0.08, 95%CI -0.10 to -0.06). For national data, an increase of 1 ?C in T corresponded to an increase in HR of 7.2 bpm (95%CI 6.2 to 8.3). After adjusting for age and gender, a 1 ?C increase in T corresponded to a mean (95%CI) 10.4 (9.5-11.4) and 6.9 (5.9-7.8) increase in HR locally and nationally, respectively.

Conclusions: Among adult ED patients nationally, for every increase in T of 1 ?C, the HR increases by approximately 7 bpm.

(C) 2019

Introduction

After the patient history, the vital signs are typically the first data points obtained and discussed by healthcare providers. These values provide crucial insight into the hemodynamic and physio- logical stability of the patient and the level of concern that the pro- vider should have. Furthermore, the constellation of vital signs in the context of the patient’s historical features paints a picture of what potential diagnoses we should consider. Careful and system- atic recording of body temperature (T) date back to at least the mid-1800s with Dr. Carl Wunderlich, who kept detailed measure- ments of his patients’ axillary temperatures and who is often given

^q Disclosures: The authors have no competing financial interests to disclose. Funding: None to report.

^qq Results from this study were presented at the SAEM Annual Meeting, May 2018, Indianapolis, IN. These results have not been previously published and are not currently under review with any other journals.

* Corresponding author at: Department of Emergency Medicine, Renaissance

School of Medicine at Stony Brook University, 101 Nicolls Rd, Stony Brook, NY 11794, United States of America.

E-mail address: [email protected] (A.J. Singer).

credit for establishing the concept of the normal body temperature range [1].

Traditional teaching is that a one degree increase in T (in degrees Celcius) should correspond to an increase in HR of approx- imately 10 beats per minute (bpm) [2]. This rule of thumb, how- ever, is derived largely from early studies conducted in children or small cohorts of otherwise Healthy adults (e.g. prisoners), which have actually found variable results depending on the age and other demographic features of subjects [3-6]. For instance, one observational study of over 60,000 children evaluated in the emer- gency department found that an increase in body temperature of 1 ?C was associated with a 10 bpm increase in HR [5]. This same study also identified a relationship between T and respiratory rate (RR) in these patients, finding a modest positive correlation, with a 1 ?C increase in body temperature associated with ~1.3 breath per minute increase in RR [5]. Meanwhile, normal ranges of pediatric vital signs vary widely with age [7], and it is unclear whether we should be extrapolating from this pediatric literature in dealing with adult patients.

Thus, the objective of this study was to determine the relation- ship between T and HR in a large number of adult emergency department (ED) patients using local and national data, controlling

https://doi.org/10.1016/j.ajem.2019.158355

0735-6757/(C) 2019

Table 1

Patient demographics, SBUH ED 2017 and NHAMCS 2015.

January 2017 SB 2015 NHAMCS

N, adult patients 6500 93.2 million

% female 54 58

Mean age (SE) 49 (0.3) 47 (0.4)

% admitted 24.8 10.8

% HR >100 18 18

Mean HR (SE) 86.4 (0.2) 86.1 (0.3)

% temp >38 1.9 1.6

Mean temperature (SE) 36.8 (0.01) 36.8 (0.01)

Mean RR (SE) 17.3 (0.1) 18.4 (0.07)

Mean SBP (SE) 136 (0.3) 138 (0.4)

N, pediatric patients 2215 30.1 million

% female 48 49

Mean age (SE) 8.0 (0.1) 7.5 (0.2)

% admitted 9.4 2.3

% HR >100 59 55

Mean HR (SE) 113.5 (0.7) 108.6 (1.0)

% temp >38 14.9 10.1

Mean temperature (SE) 37.2 (0.02) 37.1 (0.02)

Mean RR (SE) 23.1 (0.2) 23.2 (0.3)

Mean SBP (SE) 115 (0.4) 116 (0.7)

NHAMCS 2015: Total population (extrapolated) 137 million.

Population with both HR and temperature: 123.3 million (94.6 > 18 years. 28.7 <= 18).

SE, standard error.

for age and gender. We also determined the relationship between T and RR both in children and adults.

Methods

We conducted a secondary analysis of (1) electronic medical records from a large academic center (with an annual ED census of 110,000), and (2) the National Hospital Ambulatory Medical Care Survey . We included adult and pediatric patients visiting our ED in January 2017 and adult and pediatric patients in the NHAMCS 2015 database. We extracted demographic data including age, gender, as well as initial vital signs. We chose to include local data because at our institution, protocol dictates that vital signs be recorded simultaneously at triage. Although data analyzed from the NHAMCS also came from ED triage, we were uncertain as to the standardization of timing of vital sign measure- ment protocols across institutions. Thus, we elected to analyze both national and local samples, with the latter containing vital signs known to be recorded concomitantly.

For local patients, we only included patients with T-HR pairs documented within 2 min of each other. We calculated descriptive

Fig. 2. Proportion of emergency department patients who were febrile, tachycardic, or tachypneic from local and national samples.

statistics and performed linear regression to determine the rela- tionship between T and HR with and without adjustment for age and gender. Linear regression analyses were performed for the variables of T and HR, with and without adjustment for age and gender. Similar analyses were performed regarding the association of T with RR. The cut-off for statistical significance was set at p < 0.05. Data are presented as mean (95%CI) unless otherwise stated.

Results

We identified 8715 local visits to the Stony Brook University Hospital Emergency Department (SBUH ED) in the month of Jan- uary 2017, as well as 123.3 million visits (98 million in adults) to emergency departments across the nation in 2015. Demographic data for patients presenting to either SBUH ED or to EDs nation- wide (based on NHAMCS), including sex, age, percentage of patients admitted, and Physiological parameters, are displayed in the Table 1. Local and national admission rates were 25% and 11%, respectively.

Among these two groups, average age, temperature, heart rate, and respiratory rates were comparable (Fig. 1). Of the total number of patients seen in emergency departments of the two groups, a small subset had a recorded T > 38 ?C, the common threshold for defining a fever [8], and a greater proportion exhibited tachycardia (HR > 100) or tachypnea (RR > 21) (Fig. 2). In comparing those with a fever to those without, the age of those exhibiting a fever was sig-

Fig. 1. Global patient characteristics for local and national samples of emergency department visitors.

Fig. 3. Mean ages of febrile and afebrile patients from local and national samples.

Fig. 4. Mean heart rate of febrile and afebrile patients from local and national samples.

Fig. 7. Average respiratory rate of febrile and afebrile patients from local and national samples.

Fig. 5. Proportion of tachycardic, febrile and afebrile patients from local and national samples.

nificantly younger than those who were afebrile in both the local and national groups (p < 0.001, p < 0.001, respectively) (Fig. 3). For all-comers with fever, the mean HR was significantly higher than in those without fever in both the local and national groups (p < 0.001, p < 0.001, respectively) (Fig. 4). Parsing HR categorically as either tachycardic or non-tachycardic, patients with fever were significantly more likely to be tachycardic in both the local and

national groups, (p < 0.001, p < 0.001, respectively) (Fig. 5). Fig. 6 displays scatterplots of T and HR data for individual patients in the local and national groups. We observed positive correlations between T and HR for both the local and national groups (R² = 0.26, 0.16, respectively).

Considering just adult patients (age > or =18 years of age), HR increased by 10.8 bpm for every 1 ?C increase in T (95%CI, 9.9- 11.8) among patients locally, and increased by 7.2 bpm for every 1 ?C increase in T (95%CI, 6.7-7.9) among patients nationally. Of note, males consistently had lower HR than females (considering all-comers) in both the local (coefficient -1.6, 95%CI, -2.4 to

-0.8) and national (coefficient -1.9, 95%CI, -2.6 to -0.9) samples, consistent with previous findings of a stronger predominance of parasympathetic regulation on HR in female versus male adults [9]. Increasing age was associated with decreasing HR (considering all-comers) in both the local (coefficient -0.08, 95%CI, -0.1 to

-0.06) and national (coefficient -0.14, 95%CI -0.16 to -0.12) sam- ples, consistent with previous findings [10].

Among children, a greater increase in HR was observed for a given temperature change as compared to that of adults. HR increased by 21.5 bpm for every 1 ?C increase in T among the local subjects (95%CI, 20.2-22.8), and by 18.3 bpm for every 1 ?C among the national subjects (95%CI, 17.4-19.3).

We finally examined the relationship between T and RR. Parsing temperature categorically as either febrile or afebrile, we found that RR was significantly elevated in febrile patients versus afebrile patients among both the local and national subjects (p < 0.001,

Fig. 6. Scatterplots of temperature and heart rate of patients from local and national samples.

Fig. 8. Proportion of tachypneic, febrile and afebrile patients from local and national samples.

p < 0.001, respectively) (Fig. 7). Likewise, significantly more patients with fever were tachypneic as compared to patients with- out fever among both local and national subjects (p < 0.001, p < 0.001, respectively) (Fig. 8). Fig. 9 displays scatterplots of T and RR data for individual patients in the local and national groups. We observed weakly positive correlations between T and RR for both the local and national groups (R² = 0.04, 0.04, respectively). In adults, RR increased by 0.4 breaths/min for every 1 ?C increase in T among the local subjects (95%CI, -0.01-0.8), and by 0.3 for every 1 ?C among the national subjects (95%CI, 0.1-0.6). In chil- dren, RR increased by 2.7 breaths/min for every 1 ?C increase in T among the local subjects (95%CI, 2.2-3.1), and by 2.9 for every 1 ?C among the national subjects (95%CI, 2.5-3.4).

Discussion

This study provides several key novel or confirmatory insights into the relationship between vital signs in a large sample of patients presenting to EDs across the United States. Firstly, based on the national data, we found that among adults, HR can be expected to increase by ~7 bpm for every 1 ?C increase in T

observed (95%CI, 6.7-7.9). This finding is in line with prior work based on smaller samples and pediatric patients [3-6,11], although it suggests a more modest change in HR per 1 ?C increase in T as compared to previous work. Consistent with this finding, this effect appears to be magnified in children, such that larger increases in HR can be expected for every 1 ?C increase in T observed, with a

~20 bpm increase in HR for every 1 ?C increase in T (95%CI, 20.2- 22.8). After adjusting for age and gender in the national sample, HR can be expected to increase by ~7 bpm for every 1 ?C increase in T (95%CI, 5.9-7.8). Thirdly, we found that men tended to have higher HR than women and that age is associated with a decrease in HR, regardless of other hemodynamic parameters, which is con- sistent with previous work [9,10]. Of note, the trends we observed in the national data were largely reflected in the data from our local ED, indicating that our sample is likely representative of the national sample at large, at least with regard to the measured parameters.

From a physiological perspective, the findings presented in this work are unsurprising. Thermoregulation is intimately tied to the metabolic demands of the body, which govern HR and RR via hypothalamic control mechanisms. For instance, the heightened metabolic demand in febrile states (e.g., infection, thyrotoxicosis) requires an increased supply of oxygen and nutrients to the tissues, which in turn is achieved through increased pumping of deoxy- genated blood through the lungs and of oxygenated blood to the peripheral tissues [11-13]. However, the magnitude with which these parameters can be expected to change, whether within the normal temperature range or at hyperthermic temperatures, has not been heretofore systematically examined, especially among large numbers of adult patients. Together, these findings may serve to help ED providers determine what may be considered ”appro- priate” changes in HR and RR for a given change in T, ultimately guiding diagnostic and management decisions.

The present study contains a number of limitations that must be acknowledged. Firstly, the retrospective design as well as the single set of vital signs limit our ability to draw any conclusions regarding causal relationships that may or may not exist among the vital signs. Secondly, while we attempted to adjust for factors such as gender and age, a host of other potentially confounding variables, such as race, socioeconomic status, and final diagnosis, may contribute to random error in our data. As such, these data should be interpreted with caution, especially in deciding the

Fig. 9. Scatterplots of temperature and respiratory rate of patients from local and national samples.

appropriate patients in whom to apply our findings. Finally, and relatedly, the reader should keep in mind that the patients included in the study were those presenting to U.S. EDs, and so these results may not apply to patients in other settings (e.g. out- patient primary care offices, intensive care units, etc.).

In summary, we show that among adult ED patients, for every increase in 1 ?C, HR increases by ~7 bpm, while among pediatric ED patients, HR increases by ~21 bpm. Greater or lesser changes in HR may be associated with other etiologies apart from fever. These findings build on prior work examining the relationship between hemodynamic parameters, and may be useful in the work-up and management of patients presenting with Abnormal vital signs.

Declaration of Competing Interest

The authors have no relevant conflicts of interest to disclose.

References

1. Mackowiak PA, Wasserman SS, Levine MM. A critical appraisal of 98.6 degrees F, the upper limit of the normal body temperature, and other legacies of Carl Reinhold August Wunderlich. JAMA 1992;268(12):1578-80.
2. Cunha BA. The Diagnostic significance of relative bradycardia in infectious disease. Clin Microbiol Infect 2000;6(12):633-4.
3. Mackenzie J. The study of the pulse, arterial, venous, and hepatic, and of the movements of the heart. Edinburgh. R. and R. Clark, Limited; 1902.
4. Whiting M. On the association of temperature, pulse and respiration with physique and intelligence in criminals: a study in criminal anthropometry. Biometrika 1915;11:1,2):1-37.
5. Davies P, Maconochie I. The relationship between body temperature, heart rate and respiratory rate in children. Emerg Med J 2009;26(9):641-3.
6. Tanner JM. The relationships between the frequency of the heart, oral temperature and rectal temperature in man at rest. J Physiol 1951;115 (4):391-409.
7. Wallis LA, Healy M, Undy MB, Maconochie I. Age related reference ranges for respiration rate and heart rate from 4 to 16 years. Arch Dis Child 2005;90 (11):1117-21.
8. Garmel G, Mahadevan S. Fever in adults. An introduction to clinical emergency medicine. 2nd ed. Cambridge: Cambridge University Press; 2012. p. 375.
9. Kuo TB, Lin T, Yang CC, Li CL, Chen CF, Chou P. Effect of aging on gender differences in neural control of heart rate. Am J Physiol 1999;277(6):H2233-9.
10. Jose AD, Collison D. The normal range and determinants of the intrinsic heart rate in man. Cardiovasc Res 1970;4(2):160-7.
11. Tuchschmidt J, Fried J, Astiz M, Rackow E. Elevation of cardiac output and oxygen delivery improves outcome in septic shock. Chest 1992;102 (1):216-20.
12. Waldstein SS, Slodki SJ, Kaganiec GI, Bronsky D. A clinical study of thyroid storm. Ann Intern Med 1960;52(3):626-42.
13. Baracos VE, Whitmore WT, Gale R. The metabolic cost of fever. Can J Physiol Pharmacol 1987;65(6):1248-54.