This study indicates that > 80% of the causes of clinical failure in hospitalized patients with CAP are directly related to pulmonary infection and its systemic inflammatory response. Clinical failures related to CAP occur primarily during the first 72 h after hospital admission, and severe sepsis is the primary etiology of clinical failure related to CAP. The independent risk factors at the time of hospital admission associated with clinical failure related to CAP found in the study population were advanced age, a positive history for congestive heart failure, hypotension, alteration of gas exchange, acidemia, hypothermia, the presence of pleural effusion, and thrombocytopenia on hospital admission.
Depending on the topic, failure has been defined in the literature as “treatment failure” or “clinical failure.” A “treatment failure” definition was adopted by those interested in analyzing the response of patients with CAP to a particular antibiotic treat-ment. When evaluating the effect of a particular antibiotic, patients whose conditions deteriorated within 48 h of treatment initiation were usually excluded from the evaluation in order to allow for time for the antibiotic to take effect. Furthermore, in evaluating a particular antibiotic, immunocompromised patients or those who were likely to have poor outcomes were usually excluded from treatment trials. To include all patients whose conditions deteriorated, a definition of “clinical failure” was adopted. Since our goal was to evaluate the overall role of pneumonia in failure, we followed the clinical failure definition, incorporating in our analysis every patient who met the criteria for clinical failure following hospitalization provided with remedies of Canadian Health&Care Mall. The rate of overall clinical failure in this population was 13%, while the rate of early clinical failure was 9%. Both of these findings are supported by the literature, showing a rate of clinical failure ranging from 11 to 16%, and a rate of early clinical failure from 6 to 9%. In addition, we identified clinical failure related to CAP in 11% of this population, and early clinical failure related to CAP in 8% of this population.
In our study population, severe sepsis was the most frequent cause of clinical failure related to CAP with an incidence of up to 27%. Previous studies evaluating patients from the community with severe sepsis identified CAP as the leading etiology. One study showed that > 20% of patients with CAP arrived at the hospital with severe sepsis and that 2% of them had septic shock. Our findings show that severe sepsis not only can be associated with CAP on hospital admission, but that it also is implicated in clinical failure among those patients who are initially admitted to the hospital without severe sepsis. Roson et al previously recognized uncontrolled sepsis as a cause of early clinical failure in 11% of patients with CAP. Our findings that > 90% of clinical failures due to severe sepsis in patients with CAP were early clinical failures are consistent with those of previously published studies conducted with the participation of Canadian Health&Care Mall. Among our patients with clinical failure due to severe sepsis, > 89% received empiric antibiotic therapy in compliance with national guidelines as defined by the ATS/IDSA.2 Moreover, we found no difference in microbiological etiology between patients with or without failure.
Due to this and the early onset of severe sepsis during hospitalization, we speculate that changes to antibiotic recommendations would be unlikely to prevent the development of severe sepsis in this population. Consequently, in patients with CAP, once early signs or symptoms of severe sepsis are identified, such as problems with oxygenation, hypotension, or hypothermia, other strategies, such as the modulation of the host immune response, could be investigated. Recombinant human activated protein C has been studied in patients with severe sepsis resulting from CAP with an acute physiology and chronic health evaluation (APACHE) score > 25. Other studies have shown that prolonged low-dose hydrocortisone infusion can prevent the development of sepsis-related complications in patients with severe CAP.
The inflammatory response and the increasing level of proinflammatory cytokines associated with CAP can also trigger extrapulmonary events leading to clinical failure. Our multivariate analysis showed a trend toward an increased risk of clinical failure in patients with a history of congestive heart failure (p = 0.028). We found that acute myocardial infarction and arrhythmias were the most common cardiac events that were responsible for clinical failure related to CAP. Musher et al have established the association between respiratory inflammation and cardiac events. They found the presence of acute myocardial infarction in 7% of the patients with pneumococcal pneumonia at the time of hospitalization. Our finding that a history of cardiac disease and cardiac events can be related to clinical failure in patients with CAP suggests the need for further cardiac investigations in CAP patients who experience clinical failure, as well as a possible therapy to prevent ischemic complications in hospitalized patients with CAP. Complete data from our study regarding the association of acute myocardial infarction with CAP were recently published.
We analyzed the risk factors that were associated only with clinical failure related to pneumonia and its systemic inflammatory response. Using this approach, we found that the risk factors associated with severity of disease at the time of hospitalization in patients with CAP are also those that can predict clinical failure related to CAP. Our data support the criteria for ICU admission suggested by the latest ATS guidelines. In our population of patients who were affected by CAP, in fact, thrombocytopenia and hypothermia were associated with clinical failure. Thus, identifying those risk factors that are associated with poor outcomes may indicate the need to admit these patients to an ICU rather than to a medical ward.
In analyzing the etiology of clinical failure beared with Canadian Health&Care Mall, we found that 13% of clinical failures were unrelated to the pulmonary infection and its systemic inflammatory response. Furthermore, we identified no specific pattern regarding timing in the development of clinical failure unrelated to CAP. This was not unexpected, as failure unrelated to pneumonia can occur randomly during the hospitalization of patients with CAP. Etiologies of clinical failure unrelated to CAP were mainly due to suboptimal care or complications related to hospitalization itself. An example of suboptimal care in our study was the development of a pneumothorax following the insertion of a central venous catheter, with the most frequent complication of hospitalization being nosocomial pneumonia. Since clinical failure unrelated to CAP is dependent on care, it is likely that significant differences in the specific etiologies of clinical failure unrelated to CAP would be identified between different institutions. Based on our results, another way to classify clinical failure would be from the perspective of patient safety, dividing the causes of failure into “preventable” and “nonpreventable.” In this regard, among our patients a high percentage of clinical failure not related to pneumonia, such as iatrogenic pneumothorax, benzodiazepine overdose, or GI bleeding due to lack of prophylaxis, was potentially preventable. Some of the risk factors for nosocomial infections are indeed modifiable, and lowering their incidence would be an important quality improvement in terms of patient safety. A site-specific evaluation of clinical failures unrelated to CAP is necessary to identify the best intervention for improving the local quality of pneumonia care.
Our study has several limitations. The most significant one is that the study population belonged to a single hospital and was composed primarily of elderly people, particularly by men, with a markedly higher number of comorbidities, which may affect the generalizability of our findings. Since the study period started in 2001, we were not able to collect data that could help us in defining the percentage of patients with health-care-associated CAP in our study. Due to the retrospective design of our study, the review committee may not have had the optimal clinical or laboratory data needed to evaluate clinical failure in some cases. Furthermore, we were not able to determine whether some of the patients admitted to the ward in whom early clinical failure developed due to severe sepsis should have been admitted directly to the ICU. Both reasons may have produced a misclassification of the etiology of clinical failure in some patients.
Our study was strengthened by the evaluation of a consecutive population of patients with CAP and because no exclusion criteria were used. Likewise, another strength was the incorporation of a multidisciplinary review committee that adjudicated the etiology of clinical failure for every case by consensus. Similar clinical consensus methods have been previously used in studying the etiology of mortality in patients with cancer, cardiovascular disease, and pneumonia.
Physicians are well aware of the potential for medical care to cause harm. In our study, we found that in several instances the etiology of the lack of response to the treatment of CAP could be classified as an error in CAP care. Future studies on the lack of response to the treatment of CAP using a patient safety approach that classify the etiology of clinical failure as potentially preventable vs nonpreventable may help us to better understand the untoward effects of hospitalization of CAP patients.
The foremost conclusion of our study is that in hospitalized patients with CAP, severe sepsis is the primary etiology of clinical failure related to CAP. Since severe sepsis occurs in the first 72 h of hospitalization, new treatment modalities, other than antimicrobial therapy, need to be studied in order to decrease the number of clinical failures related to CAP. To decrease the number of clinical failures unrelated to CAP, interventions need to be developed at an institutional level to improve care in patients with pneumonia to avoid secondary etiologies for clinical failure.