LRTI in the Elderly: Predicting Mortality
Abstract & Commentary
Synopsis: Three clinical markers—acute exacerbation of a comorbid condition, tachypnea, and elevated CRP—can identify individuals at risk for death.
Source: Seppa Y, et al. Arch Intern Med. 2001;161:2709-2713.
Seppa and colleagues set out to answer the questions, "Is there a data set, readily available in a primary care office, that can assess the severity of lower respiratory tract infections (LRTI) in the elderly?" and "Could this data set identify individuals who needed hospitalization?"
This was a parallel study conducted in northern Finland in conjunction with a pneumococcal and influenza trial. Seppa et al began with the entire population of people older than 65 years of age in 55 municipalities. LRTI was defined as bronchitis or pneumonia and the diagnosis was made by history and physical examination. The patients were treated by primary care physicians (PCPs), who had complete discretion for decisions involving prescriptions and referral. Over the 2-year period of the study, 1743 cases of LRTI with signs suggestive of pneumonia were identified. They appropriately excluded patients with incomplete data sets or who were institutionalized, terminal, bedridden, receiving immunosuppressive drugs, or demented. The attending physicians recorded information about each patient on a case report form. This form included demographics, history of present illness, symptoms, physical findings, and comorbid conditions. The patients had some basic blood work drawn (complete blood count [CBC], C-reactive protein [CRP], and erythrocyte sedimentation rate [ESR]). As their end point, Seppa et al chose mortality due to LRTI within 30 days after the first visit to a PCP.
After exclusions, there were 950 patients with LRTI. Thirty-eight (4.1%) died within 30 days. There was no difference in mortality between men and women. A retrospective analysis of chest x-rays showed that 74% of the patients had definite or probable pneumonia, but these data were not used because less than half of the PCPs involved in this study had same-day access to chest x-rays. Only tachypnea (respiratory rate at least 25), elevated CRP (at least 100 mg/L), and acute exacerbation of a comorbidity were independent relative risk factors for mortality in a multivariate regression model. These risk factors were additive. The mortality rate was 2.2% if no more than 1 factor was present; it rose to 20% if all 3 were present.
Comment by Allan J. Wilke, MD
I have made some random observations. There are only 59,790 people older than 65 in northern Finland? And they are scattered over 55 municipalities? That is an average of just over 1000 elders per community. These are small towns! It occurred to me that they are describing the Upper Peninsula of Michigan, where I grew up. (Come to think of it, there are a lot of people of Finnish descent in the UP.) My point is, this population may or may not be similar to the one that you see. It is important to consider your population before applying the results of any study to them.
There are a lot of things to like in this study. One of them is the completeness of the study population. Another is that they were being cared for by PCPs and were not the subjects of a study conducted at a tertiary care institution. I like the simplicity of the rule and the large number of subjects.
Other prediction rules exist, but they are either considerably more complex, have been validated on populations that include others besides elderly patients or on patients who are already hospitalized, or are limited to nursing home patients. This underlines a second important point: the location of the study can produce inherent bias.
There are some weaknesses to this study. Some physicians would argue that diagnosing pneumonia by physical examination alone is antiquated with a sensitivity of 47-69% and specificity of 58-75%. If that is the case, this study is in trouble. Seppa et al discuss "LRTI with signs suggestive of pneumonia," but do not provide any information about what signs are suggestive and which are not. Many of the physicians in the study did not have same-day access to chest x-rays; this is certainly not the case here. On the other hand, CRP may not be routinely available (or at least not as quickly as one might like). ESR and CRP are both nonspecific measures of inflammation. ESR is easier, quicker, and less expensive to perform, but CRP is a better gauge of the intensity of inflammation. One of the 3 risk factors—acute exacerbation of a comorbidity—was determined by the individual attending physician and, therefore, is subject to bias.
The most glaring problem is that Seppa et al did not address their second question, "Could this data set identify individuals who needed hospitalization?" I suppose that one could assume that if a patient had no more than 1 risk factor, he could be watched as an outpatient with frequent follow-up, or if all 3 factors were present, hospitalized. Perhaps, as Seppa et al state in their final statement, "Whether these markers are generally applicable to the evaluation of the severity and treatment decisions of LRTI among elderly patients in primary care should be ascertained and validated in a new prospective study."
Dr. Wilke, Assistant Professor of Family Medicine, Medical College of Ohio, Toledo, OH, is Associate Editor of Internal Medicine Alert.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content