Fever and Infection in the Nursing Home

In her discussion on nursing home evaluations, Suzanne Bradley, MD, Associate Professor, University of Michigan and the Ann Arbor GRECC, presented symptoms and signs of infection in long-term care, an approach to diagnosis and management, useful diagnostic tests, common pathogens, and when to initiate antibiotics. Most strategies for evaluating fever and infection have been derived from hospitals, where the entire institution is devoted to the diagnosis of acute illness (ie, physicians are present, all diagnostic laboratories are onsite). In contrast, nursing homes focus more on comfort care, physicians are rarely onsite, and many diagnostic tests are not available. “So, I think the bottom line is we need different criteria to diagnose infection in long-term care,” the speaker said. Infection is common in the nursing home setting. There are also complications related to diagnosis because many patients fail to perceive symptoms, they are unable to communicate them, or the typical signs and symptoms of infection found in younger individuals may not be present in older adults because the inflammatory response may be lacking. Because physicians often rely on empiric antibiotic therapy, antibiotic resistance is common; when they fail to diagnose these processes, morbidity and mortality is also common. In general, infection rates are expected to be highest in patients with the greatest functional disability, who require the most skilled care, and who have indwelling devices. Dr. Bradley noted that infection rates and attributable mortality are probably underestimated because very rarely do physicians make an absolute diagnosis of a clinical syndrome and patients are often treated empirically. Infection rates in nursing homes are virtually identical to those in hospitals (ie, approximately 4 infections per thousand patient-days). Accounting for 90% of infections, the most common clinical syndromes are similar to those found in community-dwelling persons and those in hospital care: urinary tract and upper respiratory tract infections, and skin and soft tissue infections. Bloodstream and gastrointestinal infections are less common. Diagnosing Infection One method of diagnosing infection in the long-term care setting is the presence of an acute change in functional status (eg, increased confusion, inability to cooperate, new incontinence, falls, decreased mobility); infection most likely accounts for 77% of these symptoms. In addition, a fever of 99 degrees F or a change in baseline temperature may be very useful for detecting infection in this population. It is recommended that a lower threshold for detecting fevers be used (ie, lowering the cut-off from 101 degrees F to 99 degrees F); ability to detect infected persons increases over 80% if the threshold is lowered. Studies have demonstrated that the likelihood of having infection is greatest in nursing home patients with a temperature of 101degrees F or higher. In addition, because older adults do not mount an adequate febrile response, a temperature increase of 2.4 degrees F from baseline temperature is also a significant indication that infection is present. In addition, a respiratory rate of greater than 25 breaths per minute is a good indication that the patient is developing a lower respiratory tract infection. There is little data on other aspects of history and physical to report on. There is some variability on clinical presentation with clinical syndrome; however, if the typical signs and symptoms of infection found in hospitalized patients are present in long-term care patients, it is a good indication that the patients have the infection. On the initial examination, other predisposing factors may assist in determining where the infection is occurring. An immobile patient should be examined for pressure sores, and skin and UTIs should be considered in a patient with diabetes. Prosthetic devices such as joint replacements, pacemakers and valves, and catheters should also be assessed for the presence of infection. Foley catheters are a particular risk as bacteremic UTI increases nearly forty-fold in this population. Signs, symptoms, and laboratory evidence of dehydration may also occur with infection; prerenal azotemia and hypernatremia may be present in 60% of these patients. Testing According to the speaker, the most useful laboratory test for diagnosing infection in the nursing home is a complete blood count (CBC). Studies have shown that an increased blood count (white blood cell count > 14,000 cells/mm³) is strongly associated with the presence of infection in older adults. Urinary tract infection in the long-term care setting includes a wide variety of unpredictable organisms (eg, Escherichia coli, Pseudomonas, Enterococcus). In women, the most common bacteria is E. coli. More resistant bacteria are present such as Proteus and Providencia in addition to gram-positive cocci in the presence of obstruction, urinary stasis, or use of a urethral catheter. These urinary abnormalities are particularly common in men with UTIs. When diagnosing a UTI in long-term care, it is important to remember that 15-50% of all nursing home patients will have significant bacteriuria > 10% CFU/ml even in tha absence of symptoms; 100% in those with a long-term urinary catheter. Treatment of asymptomatic bacteriuria has no effect on patient well-being, symptoms, overall morbidity, or their mortality.   In addition, 30% of nursing home residents have significant pyuria > 10 white blood cells per high-power field. “So, positive cultures and UA are not terribly useful to differentiate true infection from those who have asymptomatic bacteriuria,” Dr. Bradley said. In the nursing home population, the diagnosis of UTI versus asymptomatic bacteriuria is made by the presence of symptoms including fever, costovertebral angle (CVA) tenderness, and new or worsening dysuria, frequency, and urgency nocturia or incontinence. Other symptoms (eg, low-grade fever, confusion, anorexia, functional decline) typically lead to evaluation and pressure from the nursing staff to treat for UTI, but more sophisticated testing (ie, urinary-antibody tests) reveal that these symptoms are rarely due to UTI. A urinalysis is useful only if patients do not have pyuria or a positive leukocyte esterase; the likelihood that they have a UTI is virtually zero. Urine culture is useful because the infecting organism and antibiotic susceptibilities are not predictable in older adults. Many organisms have been identified as causes of outbreaks of respiratory tract infection in the long-term care setting. However, in most instances, respiratory tract infections are due primarily to Streptococcus pneumoniae, Haemophilus influenza, and Moraxella catarrhalis in this population. Most long-term care facilities cannot perform blood gas analysis. It is known that a PaO₂ of < 60 is predictive of significant mortality in patients with severe community-acquired pneumonia. In nursing homes, pulse oximetry may be more useful; an oxygen saturation of less than 90%, particularly when associated with an increased respiratory rate, is highly predictive of impending respiratory failure. Pulse oximetry results may be useful in assisting the physician in making end-of-life or hospital transfer decisions. A chest x-ray still is the most reliable method for the diagnosis of pneumonia; however, it is unclear whether this diagnostic test alters outcome in nursing home patients. In patients who are suspected of having pneumonia clinically, the chest x-ray confirms the diagnosis in 75-90% of cases. Films may be useful even if the quality is poor or there is a lack of a prior film. X-ray results may alter management, (ie, an empyema might be drained, multilobular involvement portends poor prognosis leading to a change in advance directives, or presence of a mass lesion might lead to further diagnostic evaluation), require another procedure, or require transfer to a hospital. The speaker pointed out that there are no data in nursing home patients indicating that sputum gram stain improves outcome. Despite this, Dr. Bradley still recommends obtaining a sputum specimen, which may be difficult to acquire because “if the patient fails to do well, I’d like to know whether I have the wrong diagnosis or whether I’m treating with the wrong antibiotic.” When influenza is present, attack rates occur in approximately 35% of residents. Isolating patients, immunizing those who have not been immunized, and administering chemotherapy prophylaxis reduces spread, morbidity, and mortality. Decisions about antiviral prophylaxis must be made within 48 hours of diagnosis. To conduct a rapid test, a nasopharyngeal swab must be obtained and tested in the lab within 1-2 hours, which is difficult for most facilities. In addition, direct antigen testing takes hours, it is expensive, and there is poor sensitivity. Viral culture tales even longer and results are usually available in 3-7 days.  Influenza testing for surveillance purposes may be helpful, however, if you suspect cases but influenza has not yet been reported in your community by local Health Departments. Skin and soft tissue infections that occur in long-term care include scabies, epidemic conjunctivitis due to Staphylococcus aureus and Streptococcus pyogenes and adenoviruses, and cellulitis and fasciitis related to Staphylococcus aureus and Streptococcus pyogenes. In cellulitis, Group A Streptococcus and Staphylococcus aureus are the most frequent pathogens isolated. Fine needle aspirates of wounds can be done but are rarely helpful unless the patient is not responding to appropriate empiric treatment, or if there is fluctuance and you anticipate deeper incision and debridement might be necessary.  Aspiration of pus for culture and susceptibility can be helpful in directing antimicrobial therapy. Pressure ulcers are always colonized with bacteria; it is difficult to diffentiate superficial colonization of wounds from infection. Needle aspirates are low yield and technically difficult; the speaker recommended consulting with a plastic surgeon early on to get tissue, especially when the patient is not responding to a brief course of  empiric antibiotic therapy. Scabies can cause unexplained rashes. Transmission can occur from person-to-person or by fomites; clinical diagnosis can be difficult. Dr. Bradley urged against using widespread use of steroids for unexplained rashes and use of routine skin scrapings. If physicians do not have access to the appropriate equipment, a consultation with a dermatologist may be helpful. Bloodstream infection is uncommon in the long-term care setting; gram-negative bacteremia is most common. According to the speaker, only 6% of infections such as UTIs, respiratory tract infections, etc, are complicated by bloodstream infection. Bacteremic UTI occurs most commonly. Overall mortality from bloodstream infection in nursing home patients approaches 35%, especially in those with pneumonia, and most die within 24 hours of onset. There is no data that suggests that early use of blood cultures and diagnosis of bacteremia improves survival in this population; therefore, obtaining blood cultures in this population in most cases is probably not a priority, although it may be useful in some specific settings. Many organisms are associated with outbreaks of gastroenteritis. Bacterial, viral, and parasitic causes have been reported. Diarrhea may be caused by pathogens that are invasive, non-invasive, or toxin- producing. Diarrheal agents may be spread on the hands of personnel, on inanimate objects, or by contaminated food or water. C. difficile is the most likely cause of diarrhea in the long-term care facility and the most treatable. C. difficile can be suspected if a patient has been taking antibiotics in the prior 30 days and has three or more watery or unformed stools in 24 hours. Fecal leukocytes are present in less than 50% of cases and are nonspecific; it is important to obtain a stool toxin for C. difficile toxins A and B. If the stool is negative for toxins A and B, or the patient has evidence of invasive diarrhea (eg, fever or bloody diarrhea), other bacterial causes of diarrhea should be sought. Administration of Antibiotics In skin and soft tissue infection, new or increasing purulence of wounds alone is not a sufficient reason to start antibiotics. Increase in local care may be sufficient. Antibiotics should be considered if there is evidence of invasion (ie, presence of two of the following: fever, redness, tenderness, warmth, or new or increasing pain). Some caveats are that starting antibiotics alone may not be efficacious for patients who need surgical debridement for deep infection, and that other noninfectious ideologies may present similarly (eg, thromboembolic disease, gout, burns). The indication for initiation of antibiotics is strongest in treating lower respiratory tract infection in the nursing home patient with a fever of 102 degrees F and either tachypnea or a new productive cough. In patients with lower-grade fever, risk of infection is reduced and more criteria are recommended to begin therapy, such as the presence of cough and one of the following: tachycardia, respiratory rate over 25, or presence of delirium or rigors. In patients with low-grade fever and respiratory tract infection who have chronic obstructive pulmonary disease, initiation of antibiotics is recommended for new or increasing cough with purulent sputum production. In the individual without COPD and low-grade fever, initiation of antibiotics is recommended if a new cough with purulent sputum production and tachypnea or delirium is present. Other indications for initiation of antibiotics include patients who have a new infiltrate on chest x-ray associated with tachypnea, low-grade fever, or productive cough. For low-grade fever and cough alone, a chest x-ray or CBC may be useful to determine if infection is present or not. Caveats for initiation of antibiotics for lower respiratory tract infection in long-term care include lack of utility of antibiotics for influenza and other viruses, large volume aspirations within 24 hours, pulmonary emboli, or congestive heart failure. For UTIs in patients without a catheter, antibiotics are indicated in those with dysuria or the presence of fever in one of the following: new or increasing urgency, frequency, superpubic pain, hematuria, CVA tenderness, or incontinence. Patients with a chronic catheter are at much greater risk of having bacteremia related to a urinary source. In this situation, having one of the following symptoms in addition to the catheter would indicate antibiotic use: fever, CVA tenderness, rigors, or delirium. In patients with indwelling catheters, you cannot predict the organism or the antibiotic susceptibility. Dr. Bradley said that it is important to always obtain a urine culture and antimicrobial susceptibility in nursing home residents since the organism and optimal treatment are not predictable. Antibiotics do not need to be initiated urgently unless the patient looks ill, has dysuria and pain, and unless the treatment of that pain would provide some benefit to the patient. Results of urine cultures should be checked so that the patient does not remain on an inappropriate antibiotic unnecessarily.