Journal of Gerontological Nursing

CLINICAL TRIGGERS FOR DETECTION OF FEVER AND DEHYDRATION: Implications for Long-Term Care Nursing

Jean K Pals, RN, C; Andrew D Weinberg, MD, FACP; Lorraine F Beal, RN, C, MS, GNP; Paul G Levesque, RN, C, ANP; Thomas J Cunningham, PA-C; Kenneth L Minaker, MD, FRCP (C

Abstract

Fever is a common problem among long-term care (LTC) residents (Franson, Schicker, LeClair, Hoffman, & Duthie, 1988; Smith, 1982; Tresch, Simpson, & Burton, 1985). Causes of fever include infection, inflammation, or an immunologie process (Bruce & Grove, 1992), all of which can occur with surprising frequency in the elderly. Fever is sensitive to the presence of illness but does not specify the organ involved. This is a useful clinical signal for diagnostic testing and subsequent treatment (Lohmann, 1988).

Fever results from a resetting of the thermoregulatory center in the hypothalamus. During fever, endogenous pyrogens elevate the body's hypothalamic set point, resulting in temperatures within the range of 38° to 410C (Bernheim, Block, & Atkins, 1979; Castle, Norman, Yeh, Miller, & Yoshikawa, 1979). FraÜ, functionally disabled, elderly individuals are particularly susceptible to infections associated with febrile episodes (FEs) (Alvarez, Shell, Woolley, Berk, & Smith, 1988; Finkelstein, Petkun, Freedman, & Antopol, 1983; Lebow & Storer 1989; Norman, Castle, & Cantrell, 1987).

Many of the other clinical manifestations of infection and fever, such as anorexia, myalgia, malaise, weakness, and drowsiness, may be vague or nonspecific, thus making temperature (T) elevation an important clinical finding for infection. Little research has focused on the symptomatology of elderly patients when febrile (Esposito & Gleckman, 1978). Most prior investigations have been conducted with the inclusion of few elderly individuals (Bruce & Grove, 1992; Jacoby & Swartz, 1973).

Despite the high incidence of FEs among the older population and residents of LTC facilities, there is controversy among health care workers as to how best to detect fevers early. It is the general practice in most LTC facilities to obtain monthly vital signs, including T, on each resident. It is not known whether this accepted practice is helpful in detecting unsuspected fevers. Another unknown factor is what prompts the nursing staff to check a resident's T and whether it is resident-initiated or staffdetected. In ambulatory settings, it is the patient-detected fever that prompts access to needed health care.

Table

Although the presence of fever suggests infection, many elderly persons may experience infections without manifesting a significant T. A clinically significant fever has been defined differently by several authors. An arbitrary definition of fever is an oral T greater than 1010F (Larson, 1982). A moderate T from 100.5° to 1040F and a high T above 1040F (oral readings) also have been proposed (Alvarez, Shell, Woolley, Berk, & Smith, 1988). Castle and colleagues (1991) noted that nursing home residents consistently demonstrated I0F lower mean basal Ts than the widely accepted 98.60F. They also noted that 47% of FEs in this population had a "blunted" response (defined as Tmax<l01°F) to infection, although 89% of all infections did have a T response greater than 990F. Castle and colleagues (1991) have proposed that a T change of ^2.40F over the patient's basal or baseline T may be more useful as a marker for a "significant" febrile response than a specific T reading. In this population, mean T elevation during FEs was ^2.40F over all routine monthly oral/ rectal readings (^98.60F).

It is possible that a segment of the elderly nursing home population who do not appear to manifest a significant fever may be exhibiting an appropriate febrile response, obscured by their generally lower-thannormal baseline or basal body T. Therefore, it would seem clinically important that nursing staff establish a baseline T for each nursing home resident or reduce the threshold indicator for infection to at least 10O0F or 990F so that staff would be able to detect more accurately a potentially significant T elevation on each individual. This may enhance the assessment of each…

Fever is a common problem among long-term care (LTC) residents (Franson, Schicker, LeClair, Hoffman, & Duthie, 1988; Smith, 1982; Tresch, Simpson, & Burton, 1985). Causes of fever include infection, inflammation, or an immunologie process (Bruce & Grove, 1992), all of which can occur with surprising frequency in the elderly. Fever is sensitive to the presence of illness but does not specify the organ involved. This is a useful clinical signal for diagnostic testing and subsequent treatment (Lohmann, 1988).

Fever results from a resetting of the thermoregulatory center in the hypothalamus. During fever, endogenous pyrogens elevate the body's hypothalamic set point, resulting in temperatures within the range of 38° to 410C (Bernheim, Block, & Atkins, 1979; Castle, Norman, Yeh, Miller, & Yoshikawa, 1979). FraÜ, functionally disabled, elderly individuals are particularly susceptible to infections associated with febrile episodes (FEs) (Alvarez, Shell, Woolley, Berk, & Smith, 1988; Finkelstein, Petkun, Freedman, & Antopol, 1983; Lebow & Storer 1989; Norman, Castle, & Cantrell, 1987).

Many of the other clinical manifestations of infection and fever, such as anorexia, myalgia, malaise, weakness, and drowsiness, may be vague or nonspecific, thus making temperature (T) elevation an important clinical finding for infection. Little research has focused on the symptomatology of elderly patients when febrile (Esposito & Gleckman, 1978). Most prior investigations have been conducted with the inclusion of few elderly individuals (Bruce & Grove, 1992; Jacoby & Swartz, 1973).

Despite the high incidence of FEs among the older population and residents of LTC facilities, there is controversy among health care workers as to how best to detect fevers early. It is the general practice in most LTC facilities to obtain monthly vital signs, including T, on each resident. It is not known whether this accepted practice is helpful in detecting unsuspected fevers. Another unknown factor is what prompts the nursing staff to check a resident's T and whether it is resident-initiated or staffdetected. In ambulatory settings, it is the patient-detected fever that prompts access to needed health care.

Table

TABLE 1Clinical Triggers tor Temperatures Taken in 42 Residents et a Nursing Heme Care Unit

TABLE 1

Clinical Triggers tor Temperatures Taken in 42 Residents et a Nursing Heme Care Unit

The authors hypothesized that in the nursing home care unit (NHCU), the ways of detecting febrile illnesses would differ significantly. The authors further hypothesized that dehydration, as defined by biochemical means, could be detected by staffobserved changes in diet and fluid intake. The specific purpose of this study was to determine and discuss the major clinical triggers for detection of fever in the nursing home population and to examine the nursing implications of these findings.

METHODS

During a 4-month period between December, 1992, and April, 1993, a prospective study of clinically significant fevers and associated data along with therapeutic regimes was conducted. The information was gathered by a team of investigators, including a registered nurse (RN), two nurse practitioners, a physician's assistant, and a physician. The study was performed in a 130-bed NHCU located at a hospital-based Veterans Affairs Medical Center. The nursing staff at the facility were aware of the fever study, and Ts were taken at their discretion (except for the routine mandated monthly vital signs, which included a T).

For this study, a clinically significant FE was defined as a T greater (han 100.0° F oral or 101.0° F rectal for a 24-hour period, or less than 24 hours, if antibiotic therapy was begun. Ts were taken with standard mercury oral or rectal thermometers maintained in situ for 3 minutes. The sources of patient information were the nurse practitioners, physician assistant, nursing staff, patient charts, and a log book of patient problems and concerns. This log book served as the means for communication between the various health care providers and allowed an overview of all medical problems for any 24hour shift.

For all patients who were enrolled in the study, a predesigned survey form was used to record multiple clinical parameters, including triggers of fever detection. The therapeutic modalities engaged to evaluate and treat the fevers included any laboratory or radiologie testing data obtained, clinical diagnosis (source of fever), transfers to acute medicine, and outcomes. Laboratory data, when obtained, was used to make a presumptive diagnosis of dehydration, in association with the FE. This required serum sodium (Na) (abnormal values defined as 5*146 mmole/L) and blood urea nitrogen/ creatinine (BUN/Cr) ratios (abnormal values 3*25). Impaired oral intake was determined to be present if documented by staff in the progress note section of the patient's record for more than one shift during FEs.

FINDINGS

During the 4-month study, there were 48 FEs (four residents with two FEs; one resident with three FEs), with the highest recorded Ts ranging from 100.1° to 102.20F orally (mean T = 101.60F) and 101.2° to 105.3°F rectally (mean T = 1030F). Mean T elevation during FEs was ≥2.40F over all routine monthly oral/ rectal readings (^98.60F). The fevers occurred in 42 (39 men, 3 women; mean age = 75±11.3) of the 130 residents of the NHCU (37%). The primary clinical triggers noted (Table 1) were coryza (24%), feeling flushed/ warm to the touch (24%), residents' appearing ill (14%), and decline in one or more activities of daily living (ADL) (11%). Of all triggers noted for fever detection, 85% were staffdetected. Fevers and clinical triggers were detected both by licensed staff nurses and nursing assistants, although the exact percentages of detection for each were not determined. Only one fever was detected during the routine monthly vital-sign check among 520 routine monthly measurements completed during the course of this study.

The clinical diagnoses associated with these FEs are listed in Table 2. The majority of clinical reasons associated with the FEs consisted of urosepsis or urinary tract infections (UTIs), lower respiratory tract infections (including pneumonia), or coryza/viral syndromes. These three diagnostic categories accounted for 79% of all clinical diagnoses associated with FEs in this study.

Laboratory values were available for 40 FEs (Table 3). Twenty-three percent (9) had elevated BUN/Cr ratios, and 25% (10) had an elevated serum Na. In patients noted to have impaired oral intake (n = ll; 27%) as documented by the staff, increased serum Na or BUN/Cr ratios were observed in 82% (9 in 11).

DISCUSSION

Many of the residents in the study experienced multiple triggers for fever detection. However, of all the triggers noted, 85% were staffdetected versus resident-initiated. The majority of these triggers, such as coryza, feeling warm to the touch, vomiting, diarrhea, or urinary symptoms, are physical symptoms or characteristics that the nursing staff, including nursing assistants, are trained to observe. Two other triggers, that of the resident appearing ill and the change or decline in ADL, are more subtle and may require special sensitivity to detect. This finding emphasizes that among the geriatric nursing home population, many of them are unable to communicate because of physical or cognitive impairment, staff must be even more aware of subtle changes in the residents' overall functional status, assessment of ADLs, and cognition, as well as their physical condition.

Table

TABLE 2Clinical Diagnoses of 48 Febrile Episodes Among Residents of a Nursing Home Care If nit

TABLE 2

Clinical Diagnoses of 48 Febrile Episodes Among Residents of a Nursing Home Care If nit

In more than 520 monthly vitalsign measurements taken among the 130 residents, only one fever considered significant by the study criteria was detected by routine monthly surveillance. This would certainly suggest that routine surveillance of Ts is not an effective means of detecting a fever. It may, however, be helpful in establishing a resident's baseline T. Deviations from this baseline may prove useful in early detection of illness and emphasize the fact that aggressive detection and ongoing assessment on the part of the nursing staff, including nursing assistants, is needed. However, it is unclear what the frequency of routine surveillance must be to establish a reliable baseline.

Staff documentation of impaired oral intake during FEs was found to be highly associated (82%) with either increased BUN/Cr ratio or serum Na in this study. Thus, staff recognition of impaired oral intake appears to reflect dehydration reliably; it is an important clinical observation when assessing febrile residents. Because, medication profiles and renal function were not part of this study protocol, further investigation into the exact relationship between the clinical symptoms and the biochemical markers of dehydration should be completed.

The residents in this study did appear to be immunologically capable of mounting an effective febrile response (highest recorded T = 105.30F7 rectally). Whether significant illness was associated with a febrile response of 99° to 10O0F, orally or rectally in this nursing home population during the same timeframe cannot be determined from this study. However, as many as 33% of elderly persons may fail to demonstrate a febrile response (as defined by a T^sub max^≥101°F) to bacterial infections (Castle, 1991); because fever in the eiderly generally indicates the presence of a serious infection (Keating, Klimek, Levine, & Kiernan, 1984; Wasserman, Levinstein, Keller, Lee, & Yoshikawa, 1989), the absence of what staff consider to be a "significantly" elevated temperature could delay the diagnosis and treatment of an infection and potentially lead to increased mortality.

Table

TABLE 3Laboratory Data fer 4O Febrile Episodes Among Residents of a Nursing Home Care Unit

TABLE 3

Laboratory Data fer 4O Febrile Episodes Among Residents of a Nursing Home Care Unit

Although the presence of fever suggests infection, many elderly persons may experience infections without manifesting a significant T. A clinically significant fever has been defined differently by several authors. An arbitrary definition of fever is an oral T greater than 1010F (Larson, 1982). A moderate T from 100.5° to 1040F and a high T above 1040F (oral readings) also have been proposed (Alvarez, Shell, Woolley, Berk, & Smith, 1988). Castle and colleagues (1991) noted that nursing home residents consistently demonstrated I0F lower mean basal Ts than the widely accepted 98.60F. They also noted that 47% of FEs in this population had a "blunted" response (defined as Tmax<l01°F) to infection, although 89% of all infections did have a T response greater than 990F. Castle and colleagues (1991) have proposed that a T change of ^2.40F over the patient's basal or baseline T may be more useful as a marker for a "significant" febrile response than a specific T reading. In this population, mean T elevation during FEs was ^2.40F over all routine monthly oral/ rectal readings (^98.60F).

It is possible that a segment of the elderly nursing home population who do not appear to manifest a significant fever may be exhibiting an appropriate febrile response, obscured by their generally lower-thannormal baseline or basal body T. Therefore, it would seem clinically important that nursing staff establish a baseline T for each nursing home resident or reduce the threshold indicator for infection to at least 10O0F or 990F so that staff would be able to detect more accurately a potentially significant T elevation on each individual. This may enhance the assessment of each nursing home resident and may initiate earlier evaluation and treatment of a possible infectious process. Also, the practice of writing "afebrile" in notes may not be particularly helpful and should be discouraged because it gives no concrete data and may hinder the ability of subsequent nursing shifts to monitor the T change from baseline.

Although the clinical diagnosis of UTl was one of the two most common diagnoses noted among the residents in this study, the clinical trigger of "urinary symptoms" was low (less than 6%). Possible explanations for this may include chronic incontinence, residents' impaired communication of their symptoms to the staff, or masking of urinary symptoms by other triggers, such as "appearing ill" and "warm to the touch."

Dehydration remains a common fluid and electrolyte disorder among the elderly (Lavizzo-Mourey, Johnson, & Stolley, 1988) and often is associated with fever and infection (Mahowald & Himmelstein, 1981). Early identification of this problem is difficult, especially in the LTC setting. If left untreated, mortality may exceed 50% (Mahowald & Himmelstein, 1981). Frail, institutionalized elderly are at an increased risk for dehydration from a variety of causes, including decreased water intake secondary to limited access; altered sensorium; gastrointestinal disorders; fluid restriction; decreased thirst perception or increased fluid loss relating to excessive urinary loss, such as diuretic misuse, glycosuria, or diabetes insipidus; gastrointestinal losses; environmental conditions, such as during a heat wave; and chronic or acute infections (LavizzoMourey, Johnson, & Stolley, 1987).

Hypernatremia, as an indicator of dehydration due to primary freewater loss, was seen in 25% of the nursing home residents, according to this study's definition (5*146 mmole/L). Additionally, an elevated BUN/Cr ratio (5=25), another laboratory indicator of dehydration, was documented in 22.5% of the residents. The laboratory data were drawn within 48 hours of the initial T spike; thus, dehydration appears to have been present early in the course of the illness, rather than as a late complication of disease, deliberate nonintervention, or poor response to therapy.

Many of the clinical signs and symptoms of dehydration, such as dry oral mucous membranes, decreased skin turgor, orthostatic blood pressure changes, flat neck veins, or presence of constipation, as assessed by nurses, are not entirely reliable because there are age-related changes in the elderly that may confound these clinical indicators. Also, weight loss as an indicator of free-water loss may not be as useful, because it often is difficult in a LTC facility to obtain reliable daily weights. However, during dehydration and its treatment, all clinical indicators should be assessed as frequently as practicable, as important corroborative data to the concrete laboratory values obtained.

IMPLICATIONS FOR NURSES

RNs represent the largest number and percentage of professionals providing long-term care (White House Conference on Aging, Technical Committee on Health Services, 1981). In the future, nurses may assume as much as 50% or more of the outpatient care now provided to the elderly by physicians (White House Conference, 1981). Nurses increasingly will provide the major role of delivering, managing, and coordinating services for the elderly and disabled (Aiken, 1981). However, staffing problems seem to be particularly acute in nursing homes, where critically low numbers of RNs still are caring for large numbers of ill, disabled, and elderly patients (Aiken, 1981).

Based on the increasing involvement of nursing in long-term care and the high percentage of concurrent dehydration associated with febrile illness seen in this study, nursing inservice programs specific to the pathophysiology and treatment of dehydration are indicated. The ability of nursing staff to detect the clinical triggers, such as decline in ADL, residents appearing ill, or skin that is warm to the touch, associated with fever and infection may be critical to early intervention and treatment of potentially lifethreatening illness and dehydration. Because the clinical signs of dehydration, such as orthostatic hypotension, skin turgor changes, or dry mucous membranes, may not be entirely reliable in this older population, greater reliance on the nursing staff in detecting other clinical triggers will be crucial.

As elevation of T over baseline may be more critical than an absolute number; "low grade" fevers may prove to represent more serious illness in the resident than might be suspected and must be investigated. Routine T and vital signs do not appear to be useful in detecting the onset of infections.

Nurses who work in geriatrics or long-term care are generally the most appropriate - and often the only professional available - to carry out this comprehensive evaluation of the chronically ill or impaired elderly patient (Aiken, 1981; Shields & Kicks, 1982). Therefore, nurses can play an important role in the overall efforts to increase the quality and effectiveness of LTC facilities in the United States.

SUMMARY

The results of this study suggest that staff vigilance is of the utmost importance in the detection of fever of any degree. Close observation and monitoring of residents' physical condition, cognitive status, and ability to perform ADL (including feeding and oral intake for early detection of any concurrent dehydration) must be initiated and be an ongoing process of evaluation.

Although the association between alterations in oral intake and laboratory markers for dehydration seems to demonstrate that nursing staff can detect early dehydration associated with FEs, further research concerning the influence of medication and renal function must be completed to elucidate the exact relationship further.

Further research is indicated on delineating what is truly a "significant" febrile response in the nursing home setting. The LTC residents in this study clearly were able to mount an effective febrile response to infection. However, infections may exist in the absence of fever, and clinical triggers for these patients may be different, although the study protocol did not allow us to investigate this.

Routine mandated monthly vital signs in this study were found to be of little or no value in detecting fevers. The primary information to be documented from routine T recording would be a resident's basal body T, which may be useful in determining the subsequent onset of a significant illness. Further study is warranted to investigate the impact of different T-monitoring policies on the early detection of fever, potential infection, and dehydration in the LTC setting. Careful observation of the clinical status of the nursing home resident clearly is more useful than routine T monitoring.

Nursing skills in physical assessment may help augment objective laboratory parameters in evaluating institutionalized residents for the early onset of fever and associated dehydration. Staff inservices relating to the sometimes atypical clinical presentations of fever and infections among institutionalized older residents along with the pathophysiology and treatment of dehydration may increase awareness of this significant problem and may result in earlier intervention.

REFERENCES

  • Aiken, L. (1981). Nursing priorities for the 1980s; Hospitals and nursing homes. American journal of Nursing, 81, 324-330.
  • Alvarez, S., Shell, C., Woolley, T., Berk, S.L., & Smith, J-K. (1988). Nosocomial infections in long-term care facilities. Journal of Gerontology, 43, M9-M17.
  • Bernheim, H.A., Block, L.H., & Atkins, E. (1979). Fever: Pathogenesis, pathophysiology and purpose. Annals of internal Medicine, 91, 261-270.
  • Bruce, JX-, & Grove, S.K. (1982). Fever: Pathology and treatment. Critical Care Nurse, 1 2(1), 40-49.
  • Castle, M. (1979). Fever: Understanding a sinister sign. Nursing '79, 9, 27-33.
  • Castle, S-C., Norman, D.C., Yeh, M., Miller, D., & Yoshikawa, TT. (1991). Fever response in elderly nursing home residents: Are the older truly colder? Journal of the American Geriatrics Society, 39, 853-857.
  • Esposito, A.L., & Cleckman, R.A. (1978). Fever of unknown origin in the elderly. Journal of the American Geriatrics Society, 26, 498-505.
  • Finkelstein, M., Petkun, W.M., Freedman, M., & Antpol, S.C. (1983). Pneumococcal bacteremia in adults: Age-dependent differences in presentation and in outcome. Journal of the American Geriatrics Society, 31, 19-27.
  • Franson, T.R., Schicker, J.M., LeClair, S.M., Hoffman, R-G., & Duthie, E.H. (1988). Documentation and evaluation of fevers in hospital-based and community-based nursing homes. Infection Control Hospital Epidemiology, 9, 447-450.
  • Jacoby, G.A., & Swartz, M.N. (1973). Fever of undetermined origin. New England Journal of Medicine, 289, 1407-1410.
  • Rearing, H.J-, Klimek, J., Levine, D.S., & Kiernan, FJ. (1984). Effect of aging in the clinical significance of fever in ambulatory adult patients. Journal of the American Geriatrics Society, 32, 282-287.
  • Larson, E., Featherstone, H., & Petersdorf, R. (1982). Fever of undetermined origin: Diagnosis and follow-up of 105 cases, 19701980. Medicine, 61, 269-292.
  • Lavizzo-Mourey, R. (1987). Dehydration in the elderly: A short review. Journal of the National Medical Association, 79, 1033-1038.
  • Lavizzo-Mourey, R., Johnson, J., & Stolley, P. (1988). Risk factors for dehydration among elderly nursing home residents. Journal of the American Geriatrics Society, 36, 213-218.
  • Lebow, L., & Storer, P. (1989). Current concepts - Geriatrics: Care of the nursing home patient. New England Journal of Medicine, 321, 93-96.
  • Lohmann, M. (1988). Fever: Different types, different causes. Nursing '88, 18, 98-101.
  • Mahowald, J.M., & Himmelstein, D.U. (1981) Hypernatremia in the elderly: Relation to infection and mortality. Journal of the American Geriatrics Society, 29, 177-180.
  • Norman, D., Castle, S., & Cantrell, M. (1987). Infections in the nursing home. Journal of the American Geriatrics Society, 35, 796-805.
  • Shields, E-M-, & Kick, R. (1982). Nursing care in nursing homes. In L.H. Aiken (Ed.), Nursing in the 1980s: Crises, opportunities, challenges. Philadelphia: J.B. Lippincott Co., pp. 195-209.
  • Smith, LM. (1982). Infection in the elderly. Hospital Practice, 17, 69-77, 81-85.
  • Tresch, D.D., Simpson Jr., W.M., & Burton, J.R. (1985). Relationship of long-term and acute care faculties. The problem of patient transfer and continuity of care. Journal of the American Geriatrics Society, 33, 819-826.
  • Wasserman, M-, Levinstein, M., Keller, E., Lee, S-, & Yoshikawa, T.T. (1988). Utility of fever, white blood cells, and differential count in predicting bacterial infection in the elderly. Journal of the American Geriatrics Society, 37, 537-547.
  • White House Conference on Aging, Technical Committee on Health Services. Health care needs of ihe elderly. Los Angeles: Division of Research in Medical Education, University of Southern California, February 1, 1981.

TABLE 1

Clinical Triggers tor Temperatures Taken in 42 Residents et a Nursing Heme Care Unit

TABLE 2

Clinical Diagnoses of 48 Febrile Episodes Among Residents of a Nursing Home Care If nit

TABLE 3

Laboratory Data fer 4O Febrile Episodes Among Residents of a Nursing Home Care Unit

10.3928/0098-9134-19950401-04

Sign up to receive

Journal E-contents