1. Background
Despite advances in antibiotic treatments, pneumonia still has high rates of morbidity and mortality (1). So, in lower-middle-income countries, it’s the third leading cause of mortality, following ischemic heart diseases and stroke (2). Advances in antibiotic use have influenced the shift in pneumonia pathogenic patterns and led to the development of antibiotic resistance (1, 3).
Today, pneumonia, caused by multidrug-resistant (MDR) pathogens is not only limited to hospital-acquired or nosocomial infections but also can be found in patients with community-acquired pneumonia (CAP) (4). Multiple studies are performed on the infections caused by MDR pathogens in CAP patients in several hospitals of the United States and Japan (5-8). Despite the controversial findings of these studies, the most important risk factors for MDR infections include a history of antibiotic use, radiological findings of bilateral infiltrates and parapneumonic effusions, the severity of oxygenation based on the ratio of the partial pressure of oxygen in arterial blood to the fraction of inspired oxygen (PaO2/FiO2), non-ambulatory status or immobilization, and presence of one healthcare-associated pneumonia (HCAP) criterion (3, 6-8).
2. Objectives
So far, no study has evaluated the factors associated with MDR pathogenic infections in CAP patients in Indonesia. Therefore, the identification of contributing factors can help to predict MDR pathogenic infections in CAP patients.
3. Methods
This cross-sectional study was conducted from March to May 2018. In total, 155 patients were participated, 85 patients with MDR pneumonia, and 70 patients with pneumonia caused by non-MDR pathogens. Sputum of all adult patients > 18 years old with CAP was collected. To be included in the current study, the sputum samples should have the Murray and Washington’s criteria (leukocytes > 25 per low-power field and epithelial cells < 10 per low-power field). Patients were excluded if CAP was diagnosed outside before the patient being referred to Hasan Sadikin General Hospital or if commensal bacteria (Streptococcus milleri or viridans streptococci including Streptococcus mitis, Streptococcus mutans, Streptococcus oralis, Streptococcus sanguinis, and Streptococcus sobrinus) were found in the sputum culture.
MDR pathogens were defined as bacteria showing resistance to at least one antibiotic from three or more antimicrobial classes (9). The HCAP criteria were as follows being hospitalized in the past three months, dialysis, intravenous therapy in the last 30 days, and residence in a nursing home. Having a history of antibiotic use was defined as using oral or intravenous intake of antibiotics for a minimum of two days in the past 30 days (4, 7). Besides, immobilization was defined as being bedbound for three days or more (10). Microorganism identification and antibiotic resistance tests were performed using a colorimetric detection system, and turbidimetry was performed using the Vitek 2 Compact Instrument. In vitro resistance test was performed based on the Kirby-Bauer method with CLSI 2018 protocols.
Bivariate analysis was performed using the chi-square test or Fisher’s test. Variables with P values < 0.25 or theoretically important were included in the multivariate logistic regression analysis. A P value of < 0.05 was considered as statistically significant. All statistical analyses were performed in SPSS version 25 for Windows. This study was ethically approved by the Ethics Committee of Dr. Hasan Sadikin Hospital. Informed consent was taken from all participants.
4. Results
Between March and May 2018, sputum samples of 256 CAP patients were examined using the Gram staining, culturing, and sensitivity tests. The sputum samples of 69 (26.1%) patients did not meet the inclusion criteria. Also, for 11 (4.15%) patients, the sputum cultures showed no microorganism growth, and for 21 (9.05%) patients, the sputum cultures revealed commensal bacterial and fungal growth. Finally, 155 patients were included in the analysis and divided into two groups: 85 (54.8%) patients with sputum cultures showed the presence of MDR pathogens, and 70 (45.2%) patients with sputum cultures showed non-MDR pathogens.
The basic demographic characteristics of the participants are presented in Table 1. The study sample included 82 (52.9%) male and 73 (47.1%), female CAP patients. The MDR group consisted of more females (n = 45; 61.6%), while the non-MDR group consisted of more males (n = 42; 51.2%). The mean age of the participants was 54 years.
Table 1.Basic Characteristics of the Participantsa
| Basic Characteristics | Total | MDR | Non-MDR | P Value |
|---|---|---|---|---|
| Demographics | ||||
| Age | 54 ± 19 | 56 ± 19 | 51 ± 19 | 0.169 |
| Sex | 0.108 | |||
| Male | 82 (52.9) | 40 (48.8) | 42 (51.2) | |
| Female | 73 (47.1) | 45 (61.6) | 28 (38.4) | |
| Presence of one HCAP criterion | 0.047b | |||
| Yes | 48 (31.0) | 32 (66.7) | 16 (33.3) | |
| No | 107 (69.0) | 53 (49.5) | 54 (50.5) | |
| History of antibiotic use | 0.226 | |||
| Yes | 31 (20.0) | 20 (64.5) | 11 (35.5) | |
| No | 124 (80.0) | 65 (52.4) | 59 (47.6) | |
| Immobilization | < 0.001b | |||
| Yes | 24 (15.5) | 22 (91.7) | 2 (8.3) | |
| No | 131 (84.5) | 63 (48.1) | 68 (51.9) | |
| Blood gas analysis | 0.499 | |||
| PaO2/FiO2 ratio < 300 | 104 (67.1) | 59(56.7) | 45(43.3) | |
| PaO2/FiO2 ratio ≥ 300 | 51 (32.9) | 26 (51.0) | 25 (49.0) | |
| Thorax radiological findings | ||||
| Infiltrates | 0.857 | |||
| Bilateral | 63 (40.6) | 34 (54.0) | 29 (46.0) | |
| Unilateral | 92 (59.4) | 51 (55.4) | 41 (44.6) | |
| Parapneumonic effusions | 0.681 | |||
| Yes | 16 (10.3) | 8 (50.0) | 8 (50.0) | |
| No | 139 (89.7) | 77 (55.4) | 62 (44.6) |
aValues are expressed as No. (%) or mean ± SD.
bSignificant.
With 37 (23.9%) positive culture results, the Klebsiella pneumoniae was the most frequently found pathogen (Table 2). Out of 37 (45.9%) cultures, 17 were from the MDR group, while the remaining 20 (54.1%) cultures belonged to the non-MDR group. Acinetobacter baumannii was the most frequently found MDR pathogen (21.2%).
Table 2.Frequency and Percentage of Bacterial Culturesa
| Culture Results | Total (N = 155) | MDR (N = 85) | Non-MDR (N = 70) |
|---|---|---|---|
| Klebsiella pneumoniae subsp. pneumoniae | 37 (23.9) | 17 (20.0) | 20 (28.6) |
| Acinetobacter baumannii | 30 (19.4) | 18 (21.2) | 12 (17.1) |
| Pseudomonas aeruginosa | 28 (18.1) | 13 (15.3) | 15 (21.4) |
| Enterobacter cloacae | 11 (7.1) | 11 (12.9) | 0 (0.0) |
| Escherichia coli | 9 (5.8) | 8 (9.4) | 1 (1.4) |
| Staphylococcus aureus | 7 (4.5) | 2 (2.4) | 5 (7.1) |
| Staphylococcus epidermidis | 4 (2.6) | 4 (4.7) | 0 (0.0) |
| Streptococcus pneumoniae | 3 (1.9) | 0 (0.0) | 3 (4.3) |
| Achromobacterxylosoxidans | 2 (1.3) | 2 (2.4) | 0 (0.0) |
| Acinetobacter lwoffii | 2 (1.3) | 1 (1.2) | 1 (1.4) |
| Burkholderiacepacia | 2 (1.3) | 1 (1.2) | 1 (1.4) |
| Proteus mirabilis | 2 (1.3) | 1 (1.2) | 1 (1.4) |
| Pseudomonas putida | 2 (1.3) | 1 (1.2) | 1 (1.4) |
| Pseudomonas stutzeri | 2 (1.3) | 0 (0.0) | 2 (2.9) |
| Staphylococcus haemolyticus | 2 (1.3) | 2 (2.4) | 0 (0.0) |
| Achromobacterdenitrificans | 1 (0.6) | 1 (1.2) | 0 (0.0) |
| Acinetobacter spp. | 1 (0.6) | 0 (0.0) | 1 (1.4) |
| Aeromonas hydrophila | 1 (0.6) | 0 (0.0) | 1 (1.4) |
| Enterobacter aerogenes | 1 (0.6) | 1 (1.2) | 0 (0.0) |
| Enterococcus faecalis | 1 (0.6) | 1 (1.2) | 0 (0.0) |
| Enterococcus faecium | 1 (0.6) | 1 (1.2) | 0 (0.0) |
| Klebsiella oxytoca | 1 (0.6) | 0 (0.0) | 1 (1.4) |
| Klebsiella pneumoniaesubsp.ozaenae | 1 (0.6) | 0 (0.0) | 1 (1.4) |
| Raoultellaplanticola | 1 (0.6) | 0 (0.0) | 1 (1.4) |
| Staphylococcus xylosus | 1 (0.6) | 0 (0.0) | 1 (1.4) |
| Stenotrophomonas maltophilia | 1 (0.6) | 0 (0.0) | 1 (1.4) |
| Streptococcus salivarius | 1 (0.6) | 0 (0.0) | 1 (1.4) |
aValues are expressed as No. (%).
Bivariate analysis, as presented in Table 1, showed that factors associated with MDR pathogenic infections in CAP patients were immobilization status (P < 0.001) and the presence of one HCAP criterion (P = 0.047). However, a history of antibiotic use, radiological findings of bilateral infiltrates, parapneumonic effusion, and PaO2/FiO2 ratio were not significantly associated with the MDR pathogens.
Multiple logistic regression analysis, as presented in Table 3, was performed with six factors that were theoretically related to MDR pathogenic infections in CAP patients. The results showed that only immobilization status was associated with MDR pathogenic infections in CAP patients (adjusted prevalence ratio = 1.862 [1.432 - 2.420]; P < 0.001).
Table 3.Multivariate Analysis of Factors Related to MDR Pathogenic Infections in CAP Patients
| Variables | Prevalence Ratio | Adjusted Prevalence Ratio | P Value |
|---|---|---|---|
| History of antibiotic use | 1.231 (0.902 - 1.679) | 1.094 (0.744 - 1.610) | 0.647 |
| Bilateral infiltrates | 0.974 (0.727 - 1.304) | 0.833 (0.620 - 1.119) | 0.225 |
| Parapneumonic effusions | 0.903 (0.541 - 1.507) | 1.027 (0.586 - 1.802) | 0.925 |
| PaO2/FiO2 ratio < 300 | 1.113 (0.810 - 1.528) | 1.123 (0.803 - 1.571) | 0.497 |
| Immobilization | 1.906 (1.537 - 2.363) | 1.862 (1.432 - 2.420) | < 0.001a |
| Presence of one HCAP criterion | 1.346 (1.021 - 1.775) | 1.180 (0.832 - 1.674) | 0.354 |
aSignificant.
5. Discussion
As mentioned before, out of 256 CAP patients, 155 (58.5%) had positive sputum cultures. Labelle et al. (11) found that 49.5% of pneumonia patients had positive sputum cultures. This study excluded 101 patients, including 69 (26.1%) patients who could not provide a representative sputum sample, 21 (9.1%) patients whose sputum culture showed fungi and commensal bacteria, and 11 (4.15%) patients whose sputum culture showed no microorganism growth.
Non-growth in the culture may be attributed to the previous antibiotic use or inadequate sputum quality (12, 13). In the present study, the bivariate analysis showed that factors related to MDR pathogenic infections in CAP patients were immobilization status and presence of one HCAP criterion (P < 0.001 and 0.047, respectively). In patients who had a history of antibiotic use, the occurrence of MDR pathogenic infection tended to be higher (64.5%), meanwhile, the difference was not significant (P = 0.226). This finding may be due to recall bias since the data were collected from the patients’ history, not their medical records.
On the other hand, radiological findings of bilateral infiltrate, parapneumonic effusions, and PaO2/FiO2 ratio < 300 were not associated with MDR pathogenic infections. This finding is inconsistent with the results reported by Falcone et al. (8), which showed that these factors were significantly associated with MDR pathogenic infections. Plain chest radiography (CXR) is used to identify pulmonary infiltrates indicative of pneumonia and to evaluate treatment outcomes. However, it only plays a limited role in the identification of specific pathogens responsible for pneumonia, as confirmed in the current study where radiological findings were not associated with MDR pathogenic infections. Indeed, bacterial pneumonia may induce a wide range of CXR patterns (14). Moreover, several studies have described unspecific radiological findings attributed to atypical pathogens (15).
According to the multivariate analysis, the immobilization status was the strongest factor associated with the MDR pathogenic infection, with an adjusted prevalence ratio of 1.862 (1.432 - 2.420; P < 0.001). This finding is in agreement with some of the previous studies, which included immobilization status in their scoring system to predict MDR pathogenic infections in CAP patients (3, 6). Immobilization reduced the mechanical function of the lungs, which led to an increased risk of colonization and infection with respiratory tract bacteria, including Gram-negative bacteria (16, 17). This may explain why immobilization was associated with the occurrence of MDR pathogenic infections in the present study.
In the current study, the commonly found pathogens were Klebsiella pneumoniae and Acinetobacter baumannii. This finding is inconsistent with a study conducted in the United States (18), which found that Streptococcus pneumoniae was the most common pathogen. Another study conducted in some Asian countries also found that the most common pathogen was Streptococcus pneumoniae in patients with CAP (19). However, the present results are in accordance with those reported by Archana et al. from India (20) and Farida et al. (21) from Indonesia. Generally, these differences might be attributed to several factors. Unsuitable transfer of samples may reduce the isolation of more fastidious bacteria, such as Streptococcus pneumoniae (22). Also, the high temperature and humidity of West Java and Indonesia promote the growth and virulence of Gram-negative bacilli (GNB); therefore, the incidence of infections caused by GNB may be higher in tropical regions (23). It is worth mentioning that infections caused by GNB are usually more severe and require hospitalization; therefore, the prevalence of Klebsiella pneumoniae may be higher in hospitalized patients (24).
Even though Klebsiella pneumoniae was the most common pathogen in the present study, Acinetobacter baumannii accounted for the highest proportion of MDR pathogens (18/85; 21.2%). Two possible mechanisms can be used to explain these findings; upregulation of innate resistance processes by A. baumannii and capacity of A. baumannii to acquire external factors to modify its susceptibility to antibiotics. Through these mechanisms, A. baumannii is easily modified by resistance determinants and becomes resistant to several classes of antibiotics (25).
In the present study, the prevalence of MDR pathogens was 54.8%, which is lower than that reported in Ethiopia (76%) (26), but higher than the rates reported from two cities in Europe, where the prevalence rates were 3.3 and 7.6%, respectively (27). There are several possible explanations for the higher antibiotic resistance in developing countries, including inappropriate prescriptions, inadequate patient education, limited diagnostic facilities, lack of surveillance of resistance development, poor quality of available antibiotics, clinical misuse, and availability of antibiotics (28, 29).
Lack of surveillance of bacterial resistance has led to the inadequate information of health professionals in developing countries about the causative bacterial and antimicrobial patterns. Therefore, health professionals mostly rely on broad-spectrum antibiotics rather than local bacterial profiles; this practice may have led to the development of resistance. On the other hand, poor quality of available antibiotics results from the absence of appropriate regulations in antimicrobial marketing, especially in developing countries, which in turn leads to the poor quality of pharmacological agents because of factors such as storage and distribution. In fact, due to these poor conditions, drugs may be degraded and contain less than the stated dose, implying that patients consume drugs lower than the optimal dose.
Antimicrobials can also be purchased without a prescription, and the frequency of antimicrobial self-medication is high (30). Clinical misuse of antibiotics may also explain why Klebsiella pneumoniae, not Streptococcus pneumoniae (accepted globally as the most common bacteria in CAP patients), is the most common species found in the present study. In this regard, Sakeena et al. (31) reported a high rate of non-prescription sales of antibiotics in the community pharmacies of developing countries.
Amoxicillin is one of the most commonly purchased antibiotics. Similarly, Auta et al. (32) found that penicillin was the most commonly recommended and supplied medicine for the treatment of upper respiratory tract infection. These antibiotics majorly eradicate Gram-positive bacteria, such as Streptococcus pneumoniae; this may partly explain the low prevalence of Streptococcus pneumoniae in our study.
However, the current study has some limitations that should be noted. Data related to the underlying disease (including cardiovascular disease, malignancy, and neurological disorder) that could affect the etiologic of CAP were not investigated. Blood culture and serologic tests to identify other important etiologies, such as Chlamydia pneumonia, Mycoplasma pneumonia, and Legionella pneumophila, were not performed due to the limitation of resources. Another limitation of this study is the recall bias of antimicrobial use based on anamnesis. Further studies are recommended to determine other possible risk factors for MDR pathogenic infections in CAP patients.
5.1. Conclusions
Immobilization is associated with the occurrence of MDR pathogenic infections in Indonesian CAP patients. Sputum cultures with positive MDR pathogens were found in 85 (54.8%) CAP patients. In this study, Klebsiella pneumoniae was the most common pathogen, as it was found in 37 (23.9%) patients, although Acinetobacter accounted for a higher percentage of MDR pathogens (21.32%).
