J Infect Chemother xxx (2015) 1e8 Contents lists available at ScienceDirect Journal of Infection and Chemotherapy journal homepage: http://www.elsevier.com/locate/jic Original article Clinical evaluation of the need for carbapenems to treat communityacquired and healthcare-associated pneumonia Kazuhiro Kamata a, Hiromichi Suzuki a, *, Koji Kanemoto b, Yasuharu Tokuda c, Seiji Shiotani d, Yumi Hirose e, Masatsune Suzuki e, Hiroichi Ishikawa b a Division of Infectious Diseases, Department of Medicine, Tsukuba Medical Center Hospital, Tsukuba, Japan Department of Respiratory Medicine, Tsukuba Medical Center Hospital, Tsukuba, Japan Japan Community Healthcare Organization, Tokyo, Japan d Department of Radiology, Tsukuba Medical Center Hospital, Tsukuba, Japan e Department of General Medicine and Primary Care, Tsukuba Medical Center Hospital, Tsukuba, Japan b c a r t i c l e i n f o a b s t r a c t Article history: Received November 2014 Received in revised form 10 May 2015 Accepted 11 May 2015 Available online xxx Carbapenems have an overall broad antibacterial spectrum and should be protected against from the acquisition of drug resistance The clinical advantages of carbapenem in cases of pneumonia have not been certified and the need for antipseudomonal antimicrobial agents to treat healthcare-associated pneumonia (HCAP) remains controversial We introduced an antimicrobial stewardship program for carbapenem and tazobactam/piperacillin use and investigated the effects of this program on the clinical outcomes of 591 pneumonia cases that did not require intensive care unit management, mechanical ventilation or treatment with vasopressor agents [221 patients with community-acquired pneumonia (CAP) and 370 patients with HCAP] Compared with the pre-intervention period, age, comorbidities and the severity and etiology of pneumonia did not differ during the intervention period Carbapenems were rarely used during the intervention period in cases of pneumonia (CAP: 12% vs 1%, HCAP: 13% vs 1%), while antipseudomonal beta-lactam use was reduced from 33% to 8% among cases with HCAP This reduction in the rate of carbapenem administration did not have an impact on the prognosis in the cases of CAP, and the in-hospital mortality was lower among the patients with HCAP during the intervention period (15% vs 5%, p ¼ 0.013) The causes of death in the cases of HCAP were not directly related to pneumonia during the intervention period The current study shows that carbapenem use can be avoided in cases of CAP or HCAP that are not in a critical condition The frequent use of antipseudomonal betalactams does not improve the clinical outcomes of HCAP © 2015, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases Published by Elsevier Ltd All rights reserved Keywords: Healthcare-associated pneumonia Community-acquired pneumonia Carbapenem Antimicrobial stewardship Introduction Carbapenems have the broadest spectrum of activity against Gram-positive and -negative bacteria among beta-lactam antimicrobials and are currently recognized as “last-line agents” in clinical practice Since the discovery of thienamycin produced from Streptomyces cattleya [1] and the commercial implementation of treatment with imipenem/cilastatin [2], several carbapenems * Corresponding author Division of Infectious Diseases, Department of Medicine, Tsukuba Medical Center Hospital, 1-3-1 Amakubo, Tsukuba, Ibaraki 305-8558, Japan Tel.: ỵ81 (29) 851 3511; fax: ỵ81 (29) 858 2773 E-mail address: hsuzuki@tmch.or.jp (H Suzuki) (meropenem, doripenem, imipenem/cilastatin, ertapenem, biapenem, panipenem/betamipron) have been released [3], and their use has increased globally by 45% in the last 10 years [4] While carbapenems have strong potency for penicillin-binding proteins and are not hydrolyzed from many beta-lactamases, including extended-spectrum beta lactamases (ESBLs), and AmpC beta-lactamases [3], carbapenem resistance has emerged frequently during the course of clinical use of these agents [5], with the production of carbapenase and/or mutations in outer membrane porin proteins Hence, the application of these drugs should be avoided if alternative antimicrobial agents are available Pneumonia is the most common infectious disease in humans, with Streptococcus pneumoniae being the leading causative pathogen, followed by Haemophilus influenzae, Klebsiella pneumoniae, http://dx.doi.org/10.1016/j.jiac.2015.05.002 1341-321X/© 2015, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases Published by Elsevier Ltd All rights reserved Please cite this article in press as: Kamata K, et al., Clinical evaluation of the need for carbapenems to treat community-acquired and healthcareassociated pneumonia, J Infect Chemother (2015), http://dx.doi.org/10.1016/j.jiac.2015.05.002 K Kamata et al / J Infect Chemother xxx (2015) 1e8 Mycoplasma pneumoniae, Staphylococcus aureus [6e9] and others In order to identify high-risk patients with pneumonia caused by Pseudomonas aeruginosa or other multidrug-resistant organisms (MDROs), the concept of health-care associated pneumonia (HCAP) was developed, and antipseudomonal antimicrobial agents are currently recommended for patients with HCAP [10], especially those with high risks [8] However, the prognostic advantages for initial Pseudomonas coverage are unclear with respect to overtreatment, and the necessity of carbapenem administration has not been investigated in either CAP or HCAP At our facility, the frequent use of carbapenems was significant problem until June 2011, during which time antipseudomonal beta-lactams were commonly prescribed for pneumonia, especially in patients with HCAP We therefore implemented new regulations regarding our carbapenem policy [11], and since July 2011, carbapenems have been rarely used for the treatment of CAP or HCAP In this study, we investigated the need for carbapenems to treat CAP and/or HCAP among newly hospitalized adult patients Patients and methods This study was performed at Tsukuba Medical Center Hospital (TMCH, 413 beds), which is located next to the University of Tsukuba Hospital and plays a role as a tertiary emergency medical center in the Tsukuba district of Japan The intervention was performed as a part of infection control, and carbapenem use was restricted to the treatment of bacterial meningitis, febrile neutropenia, rapidly progressive sepsis, nosocomial onset intraabdominal infection and infections with highly drug-resistant Gram-negative bacteria The use of carbapenems in patients without these conditions was not recommended, and the prescribing physicians were individually instructed with daily monitoring by an Infectious Disease (ID) physician An alternative increase in the rate of tazobactam/piperacillin (TAZ/PIPC) administration was anticipated; thus, all individual prescriptions of TAZ/ PIPC were also monitored daily, and the use of these drugs in patients without life-threatening conditions, hospital-acquired infections or highly drug-resistant Gram-negative infections was not recommended during the intervention period The clinical evaluation was performed by comparing the outcomes in the pre-intervention period (July 2010 to June 2011) with those obtained during the two-year intervention period (July 2011 to June 2013) The two-year intervention period was divided into two categories, (i) phase I (July 2011 to June 2012) and (ii) phase II (July 2012 to June 2013), as most of the physicians at TMCH agreed to comply with our carbapenem policy in the phase II intervention period after being informed of the results of the phase I intervention period [11] This study was approved by the ethics committee of TMCH (approved number: 2014e002) 2.1 Patients with pneumonia and definitions of CAP and HCAP At TMCH, almost all adult pneumonia patients (18 years of age or older) who required inpatient care are admitted to the Department of Respiratory Medicine (RM) or Department of General Medicine and Primary Care (GM) Therefore, we reviewed the records of all patients discharged from RM or GM between July 2010 and June 2013 The initial chart review was performed by an ID physician (H.S.), and cases suspicious for a diagnosis of pneumonia were further reviewed by a pulmonologist (K.K) and radiologist (S.S) individually; only patients diagnosed with pneumonia by both physicians were included in this study Patients with lung abscesses [12] or empyema [13,14] and/or those diagnosed with an active infection of tuberculosis, non-tuberculous mycobacteria or fungi were excluded from this study Following the identification of adult patients clinically diagnosed with pneumonia, we reviewed their clinical information and excluded patients who did not fulfill the study criteria for pneumonia, determined based on previous research [6] Carbapenem use was not avoided in life-threatening cases and we thus excluded patients with pneumonia who required initial intensive care unit (ICU) management or treatment with vasopressor agents or mechanical ventilation in this study The requirement of ICU management was determined by each physician HCAP was considered in cases of pneumonia in which the patient met the pneumonia-specific criteria for HCAP [10], including: hospitalization for !2 days during the previous 90 days, antibiotic use during the previous 90 days, a non-ambulatory status, tube feeding, an immunocompromised status or the use of gastric acid suppressive agents 2.2 Clinical assessment and outcome measurements We compared the baseline characteristics, comorbidities, severity of pneumonia, laboratory findings on admission, causative pathogens, concurrent infections, treatment and prognosis The severity of pneumonia was assessed according to the Pneumonia Severity Index (PSI) and A-DROP scale (age, dehydration, respiratory failure, orientation disturbance and low blood pressure) [15,16] Causative pathogens were considered if (i) the bacteria were isolated from good quality sputum (Geckler class or more) and the Gram stain findings were compatible with the isolated bacteria, (ii) blood cultures were positive and the same bacteria were isolated from a sputum culture or (iii) urine antigen and/or antibody testing was positive for the target pathogens Treatment was evaluated based on the use of carbapenems, antipseudomonal agents and the frequency of combination therapy Dose adequacy was assessed according to the defined daily dose [17] adjusted for the renal function in each case The primary outcome was in-hospital mortality, and the secondary outcomes were 30-day mortality, deterioration of activities of daily living (new onset of a bed-ridden status), new requirements for tube feeding or parental nutrition (PN) and/or the new introduction of home oxygen therapy (HOT) 2.3 Statistical analysis We compared the patients with pneumonia among three periods (pre-intervention period, intervention period phase I, intervention period phase II) The comparisons were made separately between the cases of CAP and those of HCAP Categorical variables were analyzed using the c2 test, and continuous variables were assessed using a one-way analysis of variance (ANOVA) Variables found to be significantly different (p < 0.05) were further evaluated using Bonferroni-corrected P values In all patients with pneumonia in each of the three periods, independent factors associated with in-hospital mortality were assessed using a multivariable adjusted logistic regression analysis to evaluate the clinical significance of HCAP Variables with a significant association (p < 0.05) in the univariate analysis were included in the multivariate analysis after adjusting for confounding factors The SPSS version 20 software package (IBM, Armonk, NY, USA) was used for all analyses Results The hospital environment did not differ during the study period (July 2010 to June 2013) (Table 1) A total of 4313 admissions were recorded at RM or GM and assessed for eligibility (Fig 1) Consequently, we identified 591 patients with pneumonia who Please cite this article in press as: Kamata K, et al., Clinical evaluation of the need for carbapenems to treat community-acquired and healthcareassociated pneumonia, J Infect Chemother (2015), http://dx.doi.org/10.1016/j.jiac.2015.05.002 K Kamata et al / J Infect Chemother xxx (2015) 1e8 Table Details of the clinical workload of TMCH and GM/RM during the study period (July 2010eJune 2013) Pre-intervention period Number of inpatient-days in TMCH Number of ambulances received in TMCH Number of deaths during hospitalization at TMCH Proportion of carbapenem agents among total antibiotics in TMCHa Number of admissions to GM or RM (General Medicine and Primary Care) (Respiratory Medicine) Number of deaths in GM or RM (General Medicine and Primary Care) (Respiratory Medicine) Intervention period Intervention period Phase I Phase II (July 2010eJune 2011) (July 2011eJune 2012) (July 2012eJune 2013) 136421 5069 722 10.8% 1420 519 901 145 37 108 136352 5166 706 5.8% 1464 526 938 137 47 90 135236 5108 605 3.8% 1429 552 877 124 26 98 TMCH Tsukuba Medical Center Hospital, GM Department of General Medicine and Primary Care, RM Department of Respiratory Medicine a The dose of antimicrobial agents was evaluated using the ATC/defined daily dose (DDD) system [17] Fig Flowchart of the case selection process TMCH Tsukuba Medical Center Hospital, COPD chronic obstructive pulmonary disease, IP interstitial pneumonia, NTM nontuberculous mycobacteria, ICU intensive care unit fulfilled the study criteria (221 patients with CAP and 370 patients with HCAP) The subjects were divided into three categories based on the period (pre-intervention period: 182 cases, intervention period phase I: 224 cases, intervention period phase II: 185 cases) Each period included multiple episodes of pneumonia admissions, with 18 patients in the pre-intervention period, 31 patients in intervention period phase I and 15 patients in intervention period phase II The baseline characteristics of the patients with pneumonia (CAP and HCAP) in each period are described in Table Among the three periods, age was slightly higher and females were more predominant in intervention phase II The characteristics were not different between the cases of CAP and HCAP, except for age, gender, recent antibiotic use, chronic pulmonary disease and liver disease The severity of disease was not different in each group and the etiology of pneumonia was similar among the periods (Table 3) S pneumoniae was the most frequent pathogen Meanwhile, P aeruginosa was the second most frequent pathogen in the cases of HCAP, and most pneumonia cases with P aeruginosa were classified as HCAP (21/24: 88%) The details of treatment are summarized in Table In the preintervention period, carbapenems were frequently used, with the following distribution: meropenem (MEPM) ¼ 47%, doripenem (DRPM) ¼ 26%, biapenem (BIPM) ¼ 24%, panipenem/betamipron (PAPM/BP) ¼ 3% Compared with that observed in the preintervention period, the rate of initial carbapenem use was significantly reduced in the intervention period (P < 0.001), with only two cases (one of CAP, one of HCAP) initially treated with carbapenems in intervention period phase II In addition, the rate of initial use of TAZ/PIPC did not increase in the intervention period, Please cite this article in press as: Kamata K, et al., Clinical evaluation of the need for carbapenems to treat community-acquired and healthcareassociated pneumonia, J Infect Chemother (2015), http://dx.doi.org/10.1016/j.jiac.2015.05.002 K Kamata et al / J Infect Chemother xxx (2015) 1e8 Table Baseline characteristics of the 591 cases of pneumonia CAP (n ¼ 221) Preintervention Age Female Nursing home or LTCF Bed ridden Tube feeding or PN Home oxygen therapy History of aspiration Recent hospitalization Antibiotics in the past mo Antacid drugs Preceding antibiotics [Quinolones] [Others] Charlson comorbidity index Chronic pulmonary diseases Cerebrovascular diseases Heart failure Diabetes mellitus Malignancy Renal diseases Liver diseases Immunosuppression PSI score Grades 1e3 Grade Grade 5a A-DROP Grades 1e2 Grade Grade Grade 5b Bilateral pneumonia Albumin (g/dL) eGFR (mL・minÀ1・1.73 m-2) C-reactive protein (mg/dL) P Value Intervention phase I Intervention phase II n ¼ 67 n ¼ 78 n ¼ 76 77 (64.5e85) 16 (24) (2) (0) (0) (2) 10 (15) (0) (0) (0) 79.5 (68e85) 33 (42) (3) (0) (0) (4) 12 (16) (0) (0) (0) 73 (53e84.5) 24 (32) (4) (0) (0) (0) 13 (17) (0) (0) (0) (13) (13) (0e1) 22 (33) (13) 10 (15) 18 (27) (2) (3) (0) (0) 11 (14) (10) (0e1) 26 (33) 14 (18) 15 (19) 11 (14) (3) (4) (5) (0) 10 (13) 12 (16) (0e1) 23 (30) 16 (21) (11) 16 (21) (3) (5) (0) (0) 34 (51) 27 (40) (9) 28 (36) 39 (50) 11 (14) 39 (51) 25 (33) 12 (16) (12) 43 (64) 12 (18) (6) 23 (34) 3.5 (3e3.7) 43 (33e60) 11 (4e21) 12 (15) 37 (47) 26 (33) (4) 19 (24) 3.6 (3.2e3.9) 42 (30e66) (4e15) 12 (16) 38 (50) 19 (25) (9) 22 (29) 3.5 (3.1e3.8) 42 (24e65) 11 (4e20) HCAP (n ¼ 370) Preintervention 0.374 0.06 0.663 N/A N/A 0.196 0.92 N/A N/A N/A 0.903 0.736 0.91 0.489 0.317 0.161 0.879 0.783 0.024 N/A 0.156 P value Intervention phase I Intervention phase II n ¼ 115 n ¼ 146 n ¼ 109 80 (73e85) 32 (28) 23 (20) 47 (41) 20 (17) (6) 58 (50) 38 (33) 49 (43) 63 (55) 80.5 (75e87) 49 (34) 28 (19) 58 (40) 19 (13) 13 (9) 71 (49) 34 (23) 50 (34) 94 (64) 83 (78e89) 50 (46) 27 (25) 53 (49) 15 (14) (6) 53 (50) 32 (29) 29 (27) 71 (65) 11 (10) 17 (15) (1e2.5) 35 (30) 43 (37) 17 (15) 20 (17) 15 (13) 10 (9) (2) 15 (13) 16 (11) 23 (16) (1e2) 58 (40) 49 (34) 35 (24) 40 (27) 21 (14) (6) (2) 19 (13) 17 (16) 11 (10) (1e2) 26 (24) 36 (33) 24 (22) 25 (23) (8) (4) (3) (8) 21 (18) 57 (50) 37 (32) 28 (19) 69 (47) 49 (34) 18 (17) 53 (49) 38 (35) (4) 51 (44) 44 (38) 15 (13) 42 (37) 3.2 (2.7e3.6) 36 (25e56) (3e13) (3) 79 (54) 43 (30) 20 (14) 33 (23) 3.2 (2.9e3.7) 37 (25e49) (3e15) (6) 43 (39) 40 (37) 20 (18) 35 (32) 3.3 (2.9e3.6) 33 (23e45) (2e11) 0.261 0.419 0.432 0.824 0.12 0.03 0.016 0.525 0.325 0.584 0.514 0.973 0.207 0.042 0.191 0.477 0.328 0.024 0.747 0.171 0.162 0.314 0.297 0.867 0.427 0.979 0.296 0.041 0.266 0.387 0.12 All categorical data are presented as number (proportion, %) Continuous data are presented as median (interquartile range) CAP community-acquired pneumonia, HCAP healthcare-associated pneumonia, LTCF long-term care facility, PN parenteral nutrition, PSI Pneumonia Severity Index, A-DROP Japan Respiratory Society community associated-pneumonia severity index, eGFR estimated glomerular filtration rate, N/A not applicable a In the PSI, elderly patients who are nursing home residents were often classified into a higher grade due to its scoring system (e.g., a 91-year-old nursing home resident is classified into Grade if he has a history of one comorbidity and dehydration) b In our study, the cases with A-DROP grade had an initial low blood pressure (SBP < 90) without the requirement of continuous vasopressor agent use Some dehydrated patients with pneumonia had transient hypotension with an improvement of hydration after being admitted to the emergency department and antipseudomonal beta-lactams were selected in only 8% (9/109) of cases of HCAP during intervention period phase II Carbapenems were used in approximately one-fourth of the patients with HCAP (23%: 26/115), while antipseudomonal beta-lactams were used in approximately half of the patients with HCAP (43%: 49/115) in the pre-intervention period during hospitalization In contrast, the rate of use was reduced to 4% (4/109: p < 0.001) for carbapenems and 17% (18/109: p < 0.001) for antipseudomonal beta-lactams during intervention period phase II During the intervention periods, there were 39 cases (intervention period phase I: 21 cases, intervention period phase II: 18 cases) who were initially treated with antimicrobial therapy other than antipseudomonal beta-lactams and were administered antipseudomonal beta-lactams during the hospitalizations Of the 39 cases, antipseudomonal beta-lactams were administered in 22 cases (56%) due to the insufficient improvement of pneumonia, 16 cases (41%) for suspected or confirmed hospital acquired infections, such as pneumonia, and one case (3%) for suspected side effect of the initial antimicrobial therapy Of the 39 cases, only one case required mechanical ventilation and was transferred to the ICU on day This case was initially treated with ampicillin-sulbactam and a clinical improvement was achieved, however, he had rapidly deteriorating respiratory failure soon after meal recommencement The in-hospital mortality did not differ among the cases of CAP in each period Among the patients with HCAP, the mortality rate was lower in intervention period phase II than in the preintervention period [5% (5 cases) vs 15% (17 cases): P ¼ 0.013] Of the five patients, three died of aspiration following improvements in the pneumonia, one died of underlying disease (vasculitis) and one died of heart failure The 30-day mortality and rates of the new onset of a bed-ridden status or new requirements for HOT and/ or tube feeding or PN were not different between the cases of CAP and HCAP in each period For all cases in the three periods, a multivariable adjusted logistic regression analysis showed a lower serum albumin level (P < 0.001), higher PSI class (P ¼ 0.021) and the initial use of carbapenem (P ¼ 0.02) to be independently associated with the in-hospital mortality, whereas HCAP (P ¼ 0.704) was not independently related to mortality, although it was identified to be a significant factor in the univariate analysis (Table 5) Among the cases initially treated with carbapenems, an adequate dose was Please cite this article in press as: Kamata K, et al., Clinical evaluation of the need for carbapenems to treat community-acquired and healthcareassociated pneumonia, J Infect Chemother (2015), http://dx.doi.org/10.1016/j.jiac.2015.05.002 K Kamata et al / J Infect Chemother xxx (2015) 1e8 Table Etiology and concurrent infections in the 591 cases of pneumonia CAP (n ¼ 221) Preintervention Detection rate of causative pathogens Causative pathogens Mixed infections Streptococcus pneumoniae Klebsiellla pneumoniae Haemophilus influenzae Moraxella catarrhalis Legionella pneumophila Mycoplasma pneumoniae MRSA MSSA SPACE (Pseudomonas aeruginosa) Others Concurrent bacterial infectionsa Influenza Intervention phase I Intervention phase II P value HCAP (n ¼ 370) Preintervention Intervention phase I Intervention phase II P value n ¼ 67 n ¼ 78 n ¼ 76 n ¼ 115 n ¼ 146 n ¼ 109 24 (36) 31 (40) 23 (30) 0.466 41 (36) 41 (28) 42 (39) 0.183 (6) 14 (21) (2) (3) (2) (2) (5) (0) (5) (3) (2) (2) (2) (3) (5) 15 (19) (0) (5) (0) (0) 10 (13) (0) (5) (3) (3) (0) (1) (3) (0) 14 (18) (0) (3) (1) (3) (4) (0) (1) (0) (0) (0) (3) (3) 0.109 0.931 0.315 0.671 0.573 0.367 0.061 N/A 0.405 0.338 0.386 0.315 0.799 0.986 14 (12) 18 (16) (5) (3) (4) (0) (3) (4) (4) (8) (8) (6) (2) (1) 12 (8) 24 (16) (4) (2) (1) (0) (1) (3) (5) (4) (3) (2) (1) (4) 11 (10) 19 (17) (5) (5) (1) (0) (3) (3) (4) (8) (6) (5) (4) (5) 0.571 0.938 0.914 0.481 0.066 N/A 0.387 0.723 0.841 0.323 0.288 0.246 0.427 0.219 All categorical data are presented as number (proportion, %) CAP community-acquired pneumonia, HCAP healthcare-associated pneumonia, MSSA methicillin-sensitive Staphylococcus aureus, MRSA methicillin-resistant Staphylococcus aureus, SPACE Serratia marcescens, Pseudomonas aeruginosa, Acinetobacter spp., Citrobacter spp and Enterobacter spp., N/A not applicable a Concurrent bacterial infections were defined if cases with pneumonia had other infection sites or bacteremia, which were considered have different etiology Table Treatment and prognosis of the 591 cases of pneumonia CAP (n ¼ 221) Initial therapy Antipseudomonal beta-lactam [Carbapenems] [Tazobactam/Piperacillin] [Other antipseudomonal beta-lactam] Non antipseudomonal beta lactam [Third cephems] [Ampicillin/Sulbactam] Quinolones Anti-MRSA antimicrobials Others Combination therapy [Beta-lactam þ Quinolones ] [Beta-lactam þ Macrolides] Adequate dose of antibiotics Carbapenem use during hospitalization Antipseudomonal beta-lactam use during hospitalization Anti-MRSA antimicrobial use during hospitalization New incidence of becoming bed ridden New requirement for HOT New requirement for TF or PN 30-day mortality In hospital mortality P HCAP (n ¼ 370) value PreIntervention intervention phase I Preintervention Intervention phase I Intervention phase II n ¼ 67 n ¼ 78 n ¼ 76 13 (19) (12) (2) (6) 47 (70) 15 (22) 28 (42) (9) (0) (2) (9) (2) (6) 56 (84) 14 (21) 20 (30) (5) (1) (3) (1) 67 (86) 14 (18) 51 (65) (8) (0) (1) 11 (14) (8) (4) 72 (92) (5) 10 (13) (8) (1) (1) (5) 61 (80) 17 (22) 42 (55) (5) (0) (7) 15 (20) (7) (9) 67 (88) (7) 15 (20) 0.013 0.002 0.822 0.294 0.064 0.741 0.017 0.684 N/A 0.111 0.187 0.224 0.389 0.267 0.003 0.039 (0) (0) (1) 2 2 (9) (0) (3) (0) (3) (7) (3) (3) (3) (4) (10) (1) (7) (4) (5) P value Intervention phase II n ¼ 115 n ¼ 146 n ¼ 109 38 (33) 15 (13) (4) 19 (17) 75 (65) 16 (14) 56 (49) (2) (0) (0) 13 (11) (1) 11 (10) 87 (76) 26 (23) 49 (43) 20 (14) (3) (3) 11 (8) 123 (84) 31 (21) 88 (60) (1) (0) (1) 24 (16) (6) 12 (8) 132 (90) 15 (10) 35 (24) (8) (1) (3) (5) 92 (84) 26 (24) 57 (52) (6) (0) (2) (8) (2) (5) 93 (85) (4) 18 (17)