Lascarrou, Jean Baptiste MD1; Lissonde, Floriane MD2; Le Thuaut, Aurélie MSc3,4; Bachoumas, Konstantinos MD5; Colin, Gwenhael MD6; Henry Lagarrigue, Matthieu MD6; Vinatier, Isabelle MD6; Fiancette, Maud MD6; Lacherade, Jean Claude MD6; Yehia, Aihem MD6; Joret, Aurélie MD7; Lebert, Christine MD6; Bourdon, Sandra MD8; Martin Lefèvre, Laurent MD6; Reignier, Jean MD, PhD1
Patients with impaired consciousness are at high risk for macroaspiration of gastric content and subsequent aspiration pneumonia, which has been associated with acute respiratory distress syndrome (ARDS) and impaired outcome (1). The mechanisms leading to ARDS and to multiple organ failure include chemical and bacterial lung injuries related to the aspiration of both acidic gastric content and bacteria from oropharyngeal and gastric secretions (1). After aspiration, bacterial aspiration pneumonia (BAP) does not develop consistently (2, 3). However, given the potential severity of BAP and absence of reliable clinical or laboratory markers separating chemical pneumonitis from BAP, many patients in a deep coma requiring mechanical ventilation (MV) routinely receive a full course of antibiotics without prior bacterial sampling if BAP is suspected (1, 2, 4).
However, the need for antibiotic therapy has been challenged on the grounds that aspiration injury may occur without infection (5, 6). Recent guidelines emphasize the importance of minimizing unnecessary antibiotic therapy, routinely obtaining microbiologic documentation of bacterial infections, and routinely reassessing the appropriateness of antibiotic treatment after 2–3 days (7). The implementation of these guidelines can decrease mortality (8). For ventilator-associated pneumonia, restrictive strategies including routine bacterial lung sampling with discontinuation of initial antibiotics when cultures were negative were associated with lower mortality in a 2002 study (9). However, no study has evaluated a restrictive antibiotic strategy in deeply comatose patients with aspiration syndrome.
We conducted a pragmatic observational prospective study with the primary objective of estimating the frequency of BAP within 48 hours after starting invasive MV for a deep coma. We also evaluated outcomes after a restrictive antibiotic strategy in which probabilistic antibiotic treatment was confined to patients meeting predefined clinical criteria for suspected BAP; these patients routinely underwent telescopic plugged catheter (TPC) sampling and were taken off the antibiotic treatment when the samples yielded bacteria below the threshold.
MATERIALS AND METHODS
Settings and Design
We conducted a prospective observational study in a medical-surgical ICU of a French university-affiliated general hospital between November 2012 and December 2014. The study complied with the current version of the Helsinki Declaration and good clinical practice guidelines. The research project was approved by the appropriate French ethics committee (CPP Ouest III n°OBS12.01) on October 29, 2012. According to French law, informed consent was not required for this observational study of patients receiving standard care. Nevertheless, the next of kin was informed. Patients discharged from the ICU were informed of the study as soon as possible. All patients admitted to the ICU for coma who had a Glasgow Coma Scale (GCS) score less than or equal to 8 and were treated with invasive MV were considered for enrollment. Exclusion criteria were pregnancy or breastfeeding, antibiotic therapy started more than 24 hours before ICU admission, chronic neurologic disease impairing laryngeal mobility (e.g., Parkinson’s disease, Alzheimer’s disease, or amyotrophic lateral sclerosis), head and neck cancer and/or external beam radiotherapy to the laryngeal area, infection at ICU admission requiring probabilistic antibiotic treatment, refusal to participate (from the patient or next of kin), or being a correctional facility inmate.
Study Procedures and Assessments
All enrolled patients were followed prospectively and managed according to the following protocol. Suspected BAP was defined as new infiltrates on the chest radiograph with at least two of the following criteria, within 48 hours after MV initiation: body temperature greater than or equal to 38.5°C or less than or equal to 35.5°C, leukocytosis (> 10,000/mm3) or leukopenia (< 4,000/mm3), and purulent tracheobronchial aspirate. In patients receiving therapeutic hypothermia, a single criterion was sufficient to define suspected BAP (10). In patients who met the predefined criteria and were still receiving MV, TPC bronchial sampling for semiquantitative cultures was performed during bronchoscopy (11). Brief sedation was given for bronchial sampling when the patient regained consciousness before being weaned off MV. Blood for aerobic and anaerobic cultures was taken simultaneously. After sample collection, probabilistic antibiotic therapy was started. We chose to cover anaerobes, which may play a role in BAP (12), and to add clavulanic acid to cover methicillin-susceptible Staphylococcus aureus, which is commonly involved in BAP. One of two probabilistic antibiotic regimens was chosen based on the profile of predefined risk factors for antibiotic resistance, namely, age over 65 years old, institutionalization, hospital stay longer than 3 months at ICU admission, and exposure to antibiotics in the past 3 months (13). Patients with only one or none of these risk factors were given IV amoxicillin-clavulanic acid, whereas those with two or more risk factors received piperacillin plus tazobactam plus amikacin. The probabilistic antibiotic regimen was modified if needed based on the bacterial culture results. The bedside physician had the last say in choosing the probabilistic and subsequent antibiotics. Confirmed BAP was defined as a positive semiquantitative bacteriologic culture of a TPC sample obtained during bronchoscopy. The threshold for defining a positive sample was 102 cfu in the event of antibiotic therapy active on the recovered bacteria within the past 24 hours (9, 14, 15) and 103 cfu otherwise. If the sample was negative, the probabilistic antibiotic treatment was discontinued. Patients without aspiration syndrome did not receive antibiotics, even if they had a history of macroaspiration during prehospital or emergency department care. Direct examination of the TPC was performed only from 9 AM to 5 PM, that is, when the laboratory was open (16–18). The patients were classified as follows:
1) No aspiration syndrome, that is, no clinical suspicion of aspiration within 48 hours after intubation
2) Patients with aspiration syndrome: clinical suspicion of aspiration with radiologic evidence of pneumonia within the first 48 hours after intubation. Those patients were divided into the following subgroups:
a) Patients with non-bacterial aspiration pneumonitis: when the pulmonary bacteriologic sample was negative.
b) Patients with BAP: when the pulmonary bacteriologic sample was positive.
Patients with BAP received a 7-day course of antibiotics. Patients with suspected BAP after MV discontinuation did not undergo TPC sampling after extubation but received probabilistic antibiotic therapy until day 7; they were excluded from the analysis. Patients with non-bacterial aspiration syndrome were taken off probabilistic antibiotic therapy. Patients without suspected BAP did not receive antibiotics unless required by an infection at another site. Patients were monitored daily until discharge from the ICU or death.
Statistical Analysis
Qualitative variables were described as number and percentage and quantitative variables as mean ± SD if normally distributed and median (25–75th percentiles) otherwise. Between-group comparisons were performed with the chi-square test or Fisher exact test as appropriate. Numerical variables were compared using the Student t test or Wilcoxon Mann-Whitney U test depending on whether distribution was normal or not.
To identify risk factors for BAP, we performed a univariate analysis of potential associations with outcome of the following variables: leukocyte count, serum procalcitonin level, PaO2/FIO2 at ICU admission, institutionalization, witnessed aspiration, gastrointestinal tract obstruction, vomiting, GCS score, estimated time from coma onset to airway protection, Fine score, and cause of coma. All pairwise multiple comparisons were analyzed by Bonferroni’s or Dunnett’s methods. For all analyses, two-tailed p values of less than 0.05 were considered statistically significant. The sample size section is provided in the supplemental text (Supplemental Digital Content 1, http://links.lww.com/CCM/C662).
RESULTS
During the study period, 1,347 patients were screened, of whom 250 were included (eFig. 1, Supplemental Digital Content 2, http://links.lww.com/CCM/C663).
Comparison of Patients With and Without Suspected BAP
Of the 250 patients, 98 met the predefined criteria for suspected BAP (Fig. 1). The remaining 152 patients received no probabilistic antibiotics. The group with suspected BAP had a higher proportion of males than the group without suspected BAP (70.4% and 46.7%, respectively; p < 0.001) (Table 1); no other baseline characteristics differed significantly. None of the patients without suspected BAP received a diagnosis of BAP during follow-up. Compared with patients with suspected BAP, those without suspected BAP had fewer days of antibiotic treatment during the first 8 ICU days (0 [0–0] vs 6 [3–7]; p < 0.001) (Table 1). Patients without suspected BAP spent less time on MV (25.7 [13.3–72.5] vs 70.7 [34.3–154.3] hr; p < 0.001), in the ICU (49.4 [28.8–89.2] vs 107.4 [60.9–191.5] hr; p < 0.001), and in the hospital (3 [1–6] vs 6 [3–19] d; p < 0.001). Neither ICU mortality nor hospital mortality differed significantly between patients with and without suspected BAP (47 [30.9%] vs 28 [28.6%]; p = 0.69 and 48 [31.6%] vs 30 [30.6%]; p = 0.87, respectively).
Comparison of Patients With BAP to Patients With Non-Bacterial Aspiration Pneumonitis
Of the 98 patients with suspected BAP, 92 underwent bronchoscopy with TPC sampling; the remaining six patients had suspected BAP after MV discontinuation and were excluded from the analysis. No serious adverse events occurred during bronchoscopy. Cultures were positive, indicating BAP, in 43 patients, who received a full course of antibiotics. The remaining 49 patients had negative cultures and were diagnosed with non-bacterial aspiration pneumonitis (Fig. 1); among them, 33 were taken off probabilistic antibiotic therapy. Only two of these 33 patients experienced relapsing pneumonia-like symptoms, which were treated with antibiotics. One of them was weaned off invasive MV and then exhibited relapsing pneumonia-like symptoms on day 5 after ICU admission, which were treated with amoxicillin-clavulanic acid in the absence of microbiologic documentation, before discharge to a psychiatric ward on the next day. The other patient had the relapse on ICU day 4 after admission, with Pseudomonas aeruginosa in the repeat TPC sample, received piperacillin and was discharged from the ICU on day 21. The remaining 16 patients with non-bacterial aspiration pneumonitis were left on antibiotics for the following reasons: ICU discharge before the TPC result was available (n = 6), bacteremia (n = 3), other infection (community-acquired pneumonia, n = 1; meningitis, n = 1; and nosocomial pneumonia, n = 1), and undocumented severe sepsis/septic shock (n = 4).
Predictors of BAP
None of the clinical or laboratory findings within 48 hours after ICU admission was significantly associated with a subsequent diagnosis of BAP (Table 2). Including the six patients weaned off MV before the suspicion of BAP did not change the results (eTable 1, Supplemental Digital Content 1, http://links.lww.com/CCM/C662). We did not perform multivariate analysis, as no variable was associated with a p value less than 0.1 in the univariate analysis. In the 92 patients whose TPC smears were examined under the microscope, counts of erythrocytes, epithelial cells, and ciliated cells were not associated with BAP (eTable 2, Supplemental Digital Content 1, http://links.lww.com/CCM/C662), whereas a positive Gram stain was associated with BAP (p < 0.001), with a positive predictive value of 67.5% (50.9–81.4) and a negative predictive value of 69.2% (54.9–81.3).
Microorganisms Recovered From Bronchoscopy Samples
The positive bacteriologic cultures of TPC samples in 43 patients showed 66 different bacterial strains. S. aureus predominated, followed by Haemophilus influenzae and then Streptococcus pneumoniae (eTable 3, Supplemental Digital Content 1, http://links.lww.com/CCM/C662). No methicillin-resistant S. aureus strains were found. Amoxicillin-clavulanic acid was used for probabilistic antibiotic therapy in 77.6% of patients with suspected BAP and 81.6% of patients with BAP (eTable 4, Supplemental Digital Content 1, http://links.lww.com/CCM/C662).
Outcome Comparison of Patients With BAP With Patients With Non-Bacterial Aspiration Pneumonitis
The groups with non-bacterial aspiration pneumonitis and BAP did not differ significantly for MV duration (68.7 [35.4–121] vs 94.4 [43.6–167.8] hr; p = 0.13), ICU stay length (101.8 [60.3–159] vs 101.8 [60.3–159] hr; p = 0.10), hospital stay length (6 [2–19] vs 9 [3–28] d; p = 0.06), ICU mortality (p = 0.65), or hospital mortality (p = 0.38) (Table 3).
DISCUSSION
The data from this prospective observational cohort study strongly suggest the following: 1) routine antibiotic treatment in comatose patients without suspected BAP (even after macroaspiration during prehospital or emergency department care) is not warranted; 2) antibiotic treatment started in comatose patients with suspected BAP can be safely discontinued if cultures of TPC samples are negative; 3) no clinical or laboratory variables collected at ICU admission or on the day of BAP suspicion are significantly associated with BAP as opposed to non-bacterial aspiration pneumonitis; and 4) BAP is not associated with differences in MV duration, ICU and hospital stay lengths, or mortality, compared with non-bacterial aspiration pneumonitis.
Although suspected BAP is common in ICU patients (12, 19), no robust epidemiologic data are available to guide patient care in this situation (12, 20). In a cohort of 273 consecutive patients admitted to the ICU for drug overdose, 17% experienced suspected BAP, but no bacteriologic data were obtained (19). In another population with drug overdose, only 1.6% of patients had suspected BAP, but fewer than 13% of all patients required ICU admission (21). We studied only patients with coma requiring invasive MV in the ICU. These patients were monitored closely for evidence of aspiration. Suspected BAP occurred in nearly two fifths of them, but less than half of these were diagnosed with BAP and received a full course of antibiotics. This fairly high frequency rate of suspected BAP can be attributed to the severity of the acute illness, with a mean Simplified Acute Physiology Score II score of 55.7 (18.4) and a deep coma, with a mean minimal GCS score of 4 (1.6). A 2001 survey of clinical practice among members of the U.S. Society of Critical Care Medicine showed that 51.9% of physicians gave antibiotics to patients with suspected BAP without first obtaining bacteriologic documentation (4). In our study, 17.2% of patients (43/250) had bacteriologically documented BAP and 19.6% (49/250) had bacteriologically documented infection at any site. The three fifths of patients without suspected BAP received no antibiotic during the first 8 ICU days. All 98 patients with suspected BAP were given probabilistic antibiotics, which were started only after TPC sampling in 92 patients. Probabilistic antibiotics were discontinued in 33 patients with negative TPC results and no other reason for antibiotic therapy. Only two of these 33 patients subsequently experienced recurrent manifestations consistent with pneumonia. The number of days with antibiotics was lower in patients diagnosed with aspiration pneumonitis compared with those diagnosed with BAP. There is still room for decreasing the use of antibiotics, in particular by using rapid diagnostic laboratory tests.
No clinical or laboratory variables significantly predicted a diagnosis of BAP in our study. Although a positive Gram stain on the TPC sample predicted BAP as opposed to non-bacterial aspiration pneumonitis, the negative predictive value was too low to allow antibiotic discontinuation until the culture results were available. Thus, our study confirms that, in the event of suspected BAP, no criteria are available for selecting patients likely to benefit from antibiotic therapy. In addition, in an earlier study, serum procalcitonin levels failed to discriminate between BAP and non-bacterial aspiration pneumonitis (3).
Recent randomized trials evaluated whether prophylactic antibiotics decreased morbidity and mortality in patients with stroke (22), stroke and dysphagia (23), traumatic brain injury (24), or other conditions (25, 26). They consistently showed that prophylactic antibiotics decreased the risk of BAP (or “early-onset pneumonia”) in MV patients, although they failed to shorten the duration of MV (26) or to decrease mortality (22, 24–26). In our cohort, infections requiring antibiotics occurred in only 26.8% of patients (67/250), and therefore, routine prophylactic full-course antibiotic therapy would have been unnecessary in 73.2% of patients. Reserving full-course antibiotics to patients with bacteriologically documented pneumonia would decrease the risk of adverse events due to antibiotics (27). Furthermore, we found no evidence of adverse outcome associated with discontinuing probabilistic antibiotic treatment, started due to suspected BAP, when bacteriologic cultures of the TPC sample were negative. These results require further evaluation in a randomized trial.
Our study has several limitations. First, no anaerobic organisms were identified in patients with BAP, although the samples were quickly transferred, under optimal conditions, to the microbiologic laboratory. Anaerobic organisms were found less often in recent studies than in older studies (28). Furthermore, when multiple bacterial species are present, bacteriologic laboratories may fail to identify each species and may therefore miss anaerobes. However, anaerobic bacterial species are susceptible to routine antibiotics (12, 29). Second, the diagnosis of BAP or non-bacterial aspiration pneumonitis in our study was based on bacteriologic cultures of a TPC sample collected during bronchoscopy. TPC sampling can be performed without bronchoscopy. However, bronchoscopy was part of standard care in our ICU, and we maintained this practice for the current study, which met criteria for a study of standard care according to French law. Third, BAP and non-bacterial aspiration pneumonitis may coexist. However, the outcomes in our group with negative TPC cultures suggest that if non-bacterial aspiration pneumonitis was present, it resolved with a short course of antibiotics. Fourth, we did not conduct a blinded interpretation of chest x-rays, since the appropriateness of antibiotic therapy was evaluated based on clinical and laboratory findings. Last, the single-center design may limit the general applicability of our results. However, the TPC sampling technique and antibiotic treatment algorithm were very similar to those reported previously in ICU patients with drug overdose (19). Our cohort is the largest to date among studies of suspected BAP in comatose patients requiring MV. Furthermore, the patient characteristics reflect the overall population of comatose patients admitted to nonspecialized ICUs (19, 25). Finally, the prospective design strongly limits the risk of bias and allows for the conclusion that our specific protocol was associated with a decrease in antibiotic use in patients with suspected BAP although not adversely affecting morbidity or mortality. However, a randomized prospective study is needed to confirm the benefits of such a protocol.
CONCLUSIONS
Among comatose patients on MV, those without suspected BAP were not at increased risk for bacterial pneumonia when not given routine antibiotics. In those with suspected BAP, routine lung sampling and blood cultures with antibiotic discontinuation when no microorganism was found were associated with a nonsignificant decrease in antibiotic use and did not increase morbidity or mortality. Only half the patients with suspected BAP had this diagnosis confirmed by microbiologic studies.
ACKNOWLEDGMENTS
We thank A. Wolfe, MD, and A. Spiers for their assistance in preparing and reviewing the article. We are grateful to Y. Alcourt, RN; N. Maquigneau, RN; C. Rousseau, RN; A. Robert, RN; A. Deschamps, RN; V. Erragne, RN; and M. Antoine, RN for their assistance in data collection. We thank S. Martin, PharmD, for her assistance in managing the database. We thank J. Dimet, PharmD, for his assistance with the administrative process.
REFERENCES
1. DiBardino DM, Wunderink RG. Aspiration pneumonia: A review of modern trends. J Crit Care 2015; 30:40–48.
2. Lauterbach E, Voss F, Gerigk R, et al. Bacteriology of aspiration pneumonia in patients with acute coma. Intern Emerg Med 2014; 9:879–885.
3. El-Solh AA, Vora H, Knight PR 3rd, et al. Diagnostic use of serum procalcitonin levels in pulmonary aspiration syndromes. Crit Care Med 2011; 39:1251–1256.
4. Rebuck JA, Rasmussen JR, Olsen KM. Clinical aspiration-related practice patterns in the intensive care unit: A physician survey. Crit Care Med 2001; 29:2239–2244.
5. DePaso WJ. Aspiration pneumonia. Clin Chest Med 1991; 12:269–284.
6. Kwong JC, Howden BP, Charles PG. New aspirations: The debate on aspiration pneumonia treatment guidelines. Med J Aust 2011; 195:380–381.
7. Bretonnière C, Leone M, Milési C, et al; Société de Réanimation de Langue Française (SRLF); Société Française d’Anesthésie et de Réanimation (SFAR): Strategies to reduce curative antibiotic therapy in intensive care units (adult and paediatric). Intensive Care Med 2015; 41:1181–1196.
8. Garnacho-Montero J, Gutiérrez-Pizarraya A, Escoresca-Ortega A, et al. De-escalation of empirical therapy is associated with lower mortality in patients with severe sepsis and septic shock. Intensive Care Med 2014; 40:32–40.
9. Chastre J, Fagon JY. Ventilator-associated pneumonia. Am J Respir Crit Care Med 2002; 165:867–903.
10. Perbet S, Mongardon N, Dumas F, et al. Early-onset pneumonia after cardiac arrest: Characteristics, risk factors and influence on prognosis. Am J Respir Crit Care Med 2011; 184:1048–1054.
11. Pham LH, Brun-Buisson C, Legrand P, et al. Diagnosis of nosocomial pneumonia in mechanically ventilated patients. Comparison of a plugged telescoping catheter with the protected specimen brush. Am Rev Respir Dis 1991; 143(5 Pt 1):1055–1061.
12. Marik PE. Aspiration pneumonitis and aspiration pneumonia. N Engl J Med 2001; 344:665–671.
13. Ben-Ami R, Rodríguez-Baño J, Arslan H, et al. A multinational survey of risk factors for infection with extended-spectrum beta-lactamase-producing enterobacteriaceae in nonhospitalized patients. Clin Infect Dis 2009; 49:682–690.
14. Wimberley NW, Bass JB Jr, Boyd BW, et al. Use of a bronchoscopic protected catheter brush for the diagnosis of pulmonary infections. Chest 1982; 81:556–562.
15. de Lassence A, Joly-Guillou ML, Salah A, et al. Accuracy of delayed (24 hours) processing of bronchoalveolar lavage for diagnosing bacterial pneumonia. Crit Care Med 2004; 32:680–685.
16. Mertens AH, Nagler JM, Galdermans DI, et al. Quality assessment of protected specimen brush samples by microscopic cell count. Am J Respir Crit Care Med 1998; 157(4 Pt 1):1240–1243.
17. Mentec H, May-Michelangeli L, Rabbat A, et al. Blind and bronchoscopic sampling methods in suspected ventilator-associated pneumonia. A multicentre prospective study. Intensive Care Med 2004; 30:1319–1326.
18. Blot F, Raynard B, Chachaty E, et al. Value of gram stain examination of lower respiratory tract secretions for early diagnosis of nosocomial pneumonia. Am J Respir Crit Care Med 2000; 162:1731–1737.
19. Christ A, Arranto CA, Schindler C, et al. Incidence, risk factors, and outcome of aspiration pneumonitis in ICU overdose patients. Intensive Care Med 2006; 32:1423–1427.
20. Kozlow JH, Berenholtz SM, Garrett E, et al. Epidemiology and impact of aspiration pneumonia in patients undergoing surgery in Maryland, 1999–2000. Crit Care Med 2003; 31:1930–1937.
21. Isbister GK, Downes F, Sibbritt D, et al. Aspiration pneumonitis in an overdose population: Frequency, predictors, and outcomes. Crit Care Med 2004; 32:88–93.
22. Westendorp WF, Vermeij JD, Zock E, et al. The Preventive Antibiotics in Stroke Study (PASS): A pragmatic randomised open-label masked endpoint clinical trial. Lancet 2015; 385:1519–1526.
23. Kalra L, Irshad S, Hodsoll J, et al; STROKE-INF Investigators: Prophylactic antibiotics after acute stroke for reducing pneumonia in patients with dysphagia (STROKE-INF): A prospective, cluster-randomised, open-label, masked endpoint, controlled clinical trial. Lancet 2015; 386:1835–1844.
24. Sirvent JM, Torres A, El-Ebiary M, et al. Protective effect of intravenously administered cefuroxime against nosocomial pneumonia in patients with structural coma. Am J Respir Crit Care Med 1997; 155:1729–1734.
25. Vallés J, Peredo R, Burgueño MJ, et al. Efficacy of single-dose antibiotic against early-onset pneumonia in comatose patients who are ventilated. Chest 2013; 143:1219–1225.
26. Acquarolo A, Urli T, Perone G, et al. Antibiotic prophylaxis of early onset pneumonia in critically ill comatose patients. A randomized study. Intensive Care Med 2005; 31:510–516.
27. Joundi RA, Wong BM, Leis JA. Antibiotics “just-in-case” in a patient with aspiration pneumonitis. JAMA Intern Med 2015; 175:489–490.
28. Mier L, Dreyfuss D, Darchy B, et al. Is penicillin G an adequate initial treatment for aspiration pneumonia? A prospective evaluation using a protected specimen brush and quantitative cultures. Intensive Care Med 1993; 19:279–284.
29. Marik PE, Careau P. The role of anaerobes in patients with ventilator-associated pneumonia and aspiration pneumonia: A prospective study Chest 1999; 115:178–183.
critical care; inhalation; mechanical ventilation; pneumonia; pneumonitis
Supplemental Digital Content
Copyright © by 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Source
Critical Care Medicine. 45(8):1268-1275, August 2017.
Your message has been successfully sent to your colleague.
Some error has occurred while processing your request. Please try after some time.
The item(s) has been successfully added to "".
