Should blood cultures be performed in terminally Ill cancer patients?
Background: No evidence-based guidelines or protocols to treat the infection-related symptoms in cancer patients with terminal stages have been established.
Keywords: Blood culture, Respiratory tract infection, Terminally ill cancer
Terminally ill cancer patients experience a high incidence of a variety of infections. , It is thought that the majority of these patients eventually die due to infection. , Actually, every physician experiences treating infections in terminal stage cancer patients. However, no evidence-based guidelines or protocols to treat the infection-related symptoms in terminally ill cancer patients have been established. This situation makes physicians frustrated when facing such patients. Needless to say, performing blood culture before starting antibiotic therapy is very important for bacterial detection.  However, repeated venous punctures might be stressful for patients as well as the medical staff. Should blood cultures be performed for patients on best supportive care (BSC), with short life expectancy, for infection treatment? Some previously documented that antibiotic therapy was effective for terminally ill cancer patients who were symptomatic and clearly conscious in order to improve symptoms related to infection and their quality of life (QOL). ,,,, On the other hand, it is reported that antibiotic therapy is less effective among patients with respiratory tract infections than among those with urinary tract infections (UTIs). ,, It is very common that dexamethasone (DEX), palliative sedation, or both are given to terminally ill patients with lung cancer because they mostly present with severe dyspnea. ,,,,, Blood culture in the treatment for infection is a standard practice.  Nevertheless, we wish to focus on whether physicians should perform blood culture for and administer antibiotic therapy to patients who are already undergoing palliative sedation and who are unconscious. It has never been demonstrated whether a blood culture is necessary or not in the treatment of infection in these patients as far as we could determine. This retrospective study is the first report describing the therapeutic impact of blood cultures in terminally ill cancer patients. We hope that this report would be a key to establish the guidelines for the treatment of infections in terminally ill patients.
We retrospectively reviewed all the patients with terminal stage cancer who died between April 2009 and March 2010 at the Department of Pulmonology, Kameda Medical Center, Japan. The patients' characteristics, the prevalence of infection, clinical outcomes including survival time and symptoms improved after antibiotic therapy, pathogens isolated, antibiotics used, whether blood cultures and some of examinations were performed or not, and the therapeutic impact of blood culture on the management of infection treatment were evaluated. The survival time was calculated from the day of admission to death. The symptom response was reviewed from the nursing and medical records.
Comparisons of group means were made by unpaired or paired t-tests or the Mann-Whitney U-test. Contingency tables were evaluated by Fisher's exact probability test. P-values <0.05 were considered significant.
A total of 62 (44 males and 18 females) patients were eligible for this study. The median age was 73 years (range 35-98 years). Patient characteristics are shown in [Table 1]. The most common cancer was that of lung (n =59, 95.2%). Thirty-two of the 62 (51.6%) patients experienced an infection. The median survival time and median duration of antibiotic therapy were 22.6 days (SD, ±20.6; range, 1-108 days) and 6.7 days (range, 1-21 days), respectively.
Palliative medicines prescribed when patients were in their terminal stage are shown in [Table 1]. Palliative sedation therapy and DEX were offered in 31 (50%) and 24 (38.7%) patients, respectively. The comparison of palliative medicines prescribed to patients with and without infection is shown in [Table 2]. Palliative sedation therapy and DEX were used in both groups. Nonsteroidal anti-inflammatory drugs (NSAIDs) were used more frequently in patients with infection than those without (P=0.04) as shown in [Table 2].
Site of infection and antibiotics prescribed
The distribution of infections by site and antibiotics prescribed is presented in [Table 3]. The most frequent site of infection was respiratory tract (n=27, 84.4%), followed by urinary tract (n=4, 12.5%). Two cases had pneumonia complicated with a UTI. Antibiotic therapies were performed in 30 of the 32 patients with infection (93.8%). The most common antibiotics prescribed were tazobactam/piperacillin (TAZ/PIPC) alone (n=16, 53.3%), followed by ceftriaxone (CTRX) alone (n=4, 13.3%).
Blood cultures were performed in 28 patients with infection (n=32, 87.5%). Sputum and urine cultures were performed in 22 (68.8%) and 12 (37.5%) patients, respectively. Four cases (n=28, 14.3%) showed positive blood cultures. Among all the courses of infection, the most common pathogens isolated were Staphylococcus aureus (n=32, 12.5%) and Escherichia More Details coli (n=32, 12.5%), followed by Haemophilus influenzae and Klebsiella pneumonia [Table 4]. Of the four cases who showed positive blood cultures, two had UTI, one had cholangitis, and one had pneumonia.
As for clinical courses, 3 of the 30 patients who received antibiotic therapy survived longer than 2 weeks. On the other hand, 23 of them died within 1 week after antibiotic therapy was started. Patients' symptom responses are shown in [Table 5]. Three of them experienced improved symptoms after antibiotic therapy. Twenty-seven patients were not confirmed as having any symptom improvement. Fever was improved in 12 (52.2%) and 14 (77.8%) cases in patients with and without infection, respectively. Other symptoms except for fever were not confirmed to show any improvement.
There were no significant differences about symptom improvement after antibiotic therapy and/or palliative medicine in both groups [Table 5].
The comparison of median survival times of the three groups is shown in [Table 6]. There were no significant differences in median survival times in these groups.
Therapeutic impact of blood culture
Four (14.3%) positive cultures were confirmed in blood obtained from 28 individual patients. Two of the four cases, one case with UTI and the other one with cholangitis complicated with UTI, had a change in the therapeutic regimen due to antibiotic susceptibility.
Malignancy is a leading cause of death worldwide, and accounted for 7.6 million deaths (around 13% of all deaths) in 2008. Deaths from cancer are projected to continue to rise to over 11 million in 2030, and lung cancer is the leading cause of cancer death worldwide.  Terminally ill patients with lung cancer are commonly complicated with respiratory tract infections. Physicians often have difficulty in treating infections in terminally ill cancer patients because there are no evidence-based protocols or guidelines. One-third of people die from cancer in Japan according to National Vital Statistics (2009-2010).  It is time that these clinical protocols or guidelines should be established.
We analyzed the necessity of antibiotics and blood cultures for infection in terminally ill cancer patients in this study. The therapeutic indication of blood cultures affected two cases (n=42, 4.4%) as described above. Both of them, which were diagnosed as UTI and cholangitis, led to a change in antibiotics according to the result of pathogens isolated. Although a blood culture is one of the most sensitive methods for the detection of bacteremia, sepsis, or infectious endocarditis,  the therapeutic impact of this examination is quite limited. Some clinicians previously reported that blood culture was not necessary for patients with both community-acquired ,, and hospital-acquired pneumonia , because its therapeutic impacts were limited. In addition, repeated venous punctures while performing blood cultures would be stressful for patients as well as the medical staff as mentioned above. Nakagawa et al. demonstrated that antibiotic therapy improved patients' QOL, showing effectiveness for terminally ill patients.  We found that 2 of 30 cases who received antibiotic therapy (one with UTI, and the other with cholangitis) had improved symptoms related to infections as described above. In contrast, in the remaining 28 cases, blood cultures could not contribute to the treatment because all of them were on palliative sedation or were unconscious due to CO 2 narcosis. Ninety-eight percent of patients in our study had lung cancer. This is why our result did not show antibiotic therapy as effective among terminally ill cancer patients.
It is striking that 23 of the 30 (76.7%) patients who received antibiotic therapy died within 1 week after the therapy was begun, and 3 lived longer than 3 weeks. However, we found that none of them experienced improved symptoms. Mostly, terminally ill patients with lung cancer present with severe dyspnea, expectoration of much sputum, and sticky cough, resulting in palliative sedation therapy which is offered to them to ease their symptoms. ,, NSAIDs and DEX were commonly used for these patients in providing palliative care. It might be impossible to properly evaluate the efficacy of antibiotic therapy in the cancer patients by fever or laboratory indicators of an inflammation, such as CRP, ESR, etc., which could be masked by palliative medicines. In the case of terminally ill patients who are unconscious, we were wondering if antibiotics as well as blood cultures were necessary. Most physicians tend to hesitate to withhold antibiotic therapy for any infectious disease, even when patients are unconscious with a terminally staging cancer. Blood cultures or antibiotic therapy for terminally ill cancer patients who would not live more than 1 week might be without merit and only result in the increase of the medical cost as well as medical staff's labor. In Japan, discussing this kind of issue has been avoided. We are not saying that all antibiotic therapies for terminally ill cancer patients with infection are not useful. Clinicians must consider whether antibiotics should be prescribed or blood cultures should be performed based on patients' life expectancy and conscious state. We strongly believe that prescribing morphine, DEX, and NSAIDs for patients with infection, especially for those with respiratory infection, without performing blood cultures or antibiotic administration might be acceptable for symptom relief and saving medical staff's labor.
As for antibiotic therapy and bacteriological findings, although not all the pathogens isolated were considered as causative, they were within the coverage of empiric therapy. This illustrates that empiric antibiotic therapy would be sufficient and reasonable in terminally ill cancer patients for infection treatment without performing blood cultures, when necessary. We suggest that not using antibiotics but administering DEX, intravenous analgesics such as NSAIDs added to palliative sedation, would be one of the useful alternatives for those patients. In getting advance directives, it would be helpful to discuss with the patients and their family whether antibiotic therapy and a blood culture should be performed when there will be little or no merit.
In terms of medical cost savings, it is roughly estimated to save US $300 by eliminating each set of blood cultures and US $1470-1960 per patient by eliminating a 2-week-long antibiotic therapy regimen.
The limitation of our study is that it is a retrospective analysis based on a very small population. Also, we did not use an objective tool to evaluate the symptoms of terminally ill cancer patients. A prospective study using an objective tool will be necessary, and more cases are to be expected.
In conclusion, blood cultures have a limited value in terminally ill patients admitted in a palliative care unit. Antibiotic therapy might not be effective in improving symptoms in terminally ill patients with lung cancer. We suggest that not performing a blood culture and withholding an antibiotic therapy would be one choice for those with respiratory tract infections if patients' life expectancy is very short and there is no perceived clinical or QOL merit.
We are grateful for the diligent and thorough critical reading of our manuscript by Mr. John Wocher, Executive Vice President and Director, International Affairs/International Patient Services, Kameda Medical Center (Japan), and Mr. Michael Burke, Summer Intern from the University of Iowa at Kameda Medical Center.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]