Overall, 50 hospitals (39.7% of Greek hospitals) were included from 12 out of 13 NUTS-2 level country regions; 25 (50%) secondary hospitals, 21 (42%) tertiary and 4 (8%) paediatric. Two hospitals denied participation and were replaced. A total of 9,707 hospitalised patients were included in the analysis.

Among the 9,707 inpatients, 1,175 had at least one HAI resulting in a prevalence of 12.1%. HAI prevalence was 14.3% in tertiary hospitals, 7.5% in secondary hospitals and 3.5% in paediatric hospitals. On average, 1.2 infections (range 1 to 4) were reported for each infected patient.

The highest prevalence was recorded in intensive care units (ICU) where 45.7% of patients had at least one HAI. Medical and surgical specialties followed with a prevalence of 13.5% and 8.2%, respectively. HAI prevalence was lower among patients in paediatric and neonatal specialties (3%), obstetrics and gynaecology (1.4%), and psychiatry (1.2%).

Of the 1,408 HAIs, 69.6% (n = 980) occurred during current hospitalisation, 28.6% (n = 403) were already present on admission and in 1.8% (n = 25) the origin was unknown. Among the infections present on admission, 26.3% (n = 106) originated from the same hospital (the patient was readmitted), 31.2% (n = 126) from another hospital, 10.1% (n = 41) from LTCFs and the rest had an unspecified origin.

Lower respiratory tract infections (LRTI) accounted for 28.9% (n = 407) of all HAIs, most of which were cases of pneumonia (n = 351, 24.9% of total infections). Bloodstream infections (BSI) were the second most reported HAI type (n = 282, 20%), among which 4.6% (n = 65) were catheter-related. Urinary tract infections (UTI) followed (n = 184, 13.1%) (Table 1).

SARS-CoV-2 infections accounted for 8.4% of HAIs (n = 118) and most of them were assessed as of mild or moderate severity (n = 61, 51.7%) followed by increased severity (n = 48, 40.7% of SARS-CoV-2 infections), with asymptomatic infections being the least reported (n = 9, 7.6%).

Device-associated infections varied by HAI type. A vascular catheter was present in 61.6% of BSI that reported device presence (n = 172/279), intubation was reported in 38.3% (n = 129/337) of pneumonia cases and a urinary catheter in 77.1% (n = 135/175) of UTIs.

In 58.9% (n = 830/1,408) of HAI at least one microorganism was identified (in total, 1,259 isolated microorganisms). Bacteria comprised most of the identified pathogens (n = 1,053, 83.6%), followed by fungi (n = 115, 9.1%) and viruses (n = 91, 7.2%).

Gram-negative bacteria were most often reported (n = 735, 58.4% of total isolates). Bacterial isolates belonged most often to Klebsiella spp. (n = 258, 20.5% of all isolates), Acinetobacter spp. (n = 161, 12.8%) and P. aeruginosa (n = 128, 10.2%). Gram-positive bacteria accounted for 24.9% of isolates (n = 313), commonly identified as Staphylococcus aureus (n = 79, 6.3%), Enterococcus spp. (n = 78, 6.2%) and coagulase negative Staphylococci (n = 77, 6.1%). Among fungi, Candida spp. (n = 99, 7.9%) were most frequently identified.

Susceptibility results were available in 93.9% of bacterial isolates. Results for specific markers were available in 95.1% of them included in the composite AMR index and in 93.4% of Enterobacteriaceae. The AMR composite index was 69.3%. Enterobacteriaceae were found to be non-susceptible to carbapenems in 46.8% (Table 2).

AU prevalence was overall 55.4% (n = 5,376/9,707); 64.9% in tertiary, 53.5% in secondary and 39.5% in paediatric hospitals. Each patient on antimicrobial therapy received on average 1.7 antimicrobials. Among treated patients, 51.9% (n = 2,788) received one antimicrobial, 35.1% (n = 1,890) received two and 13% (n = 698) received at least three.

AU prevalence was highest in ICU in 71.3% (n = 429/602). Among patients of surgical and medical specialties, AU prevalence was 59.4% (n = 1,781/2,999) and 58.2% (2,649/4,548), respectively.

Overall, 9,003 antimicrobials were recorded. Antimicrobial agents belonged most frequently to the class of penicillin combinations including beta-lactamase inhibitors in 16.2% (n = 1,460) and followed by second-generation cephalosporins in 10.9% (n = 978), glycopeptides in 9.4% (n = 847), fluoroquinolones in 9.2% (n = 827) and carbapenems in 8.3% (n = 748). Most frequently recorded antimicrobial agents, accounting for 75% of total AU are presented in Fig. 1.

Antimicrobial agents (ATC5 code) accounting for 75% of study antimicrobial use

Antimicrobials were intended for treatment of community-acquired infections in 34.6% of all prescriptions (n = 3,114), hospital-acquired infections in 22.9% (n = 2,062) and LTCF-acquired infections in 5.1% (n = 460). Prophylactic use followed with 10.9% (n = 983) prescribed for medical conditions and 18.4% (n = 1,656) for surgical procedures. The indication was other or unknown in 5.6% (n = 502) and 2.5% (n = 226) of antimicrobials, respectively.

Diagnosis for antimicrobials prescribed for treatment was provided in 97.9% of the cases (n = 5,520/5,636). The most frequent antimicrobial classes (75% of total) by infection origin and reported diagnoses are presented in Fig. 2. Absolute frequencies are shown for the most common (75%) ones in each class and the rest of the diagnoses are grouped in the category “other”. The most frequent antimicrobial classes were carbapenems for treatment of hospital infections accounting for 14.3% (n = 289/2,025) and beta lactam combinations for community-acquired infections in 19.9% (606/3,039).

The 75% most frequent antimicrobial classes (ATC4 code) indicated for treatment of Community-acquired (A) and Hospital-acquired (B) infections, by most commonly reported diagnoses. Values below 5 are not shown

As for surgical prophylaxis prescriptions (n = 1,656), a single dose was given in 5.8% of cases (n = 96), one-day courses in 18.5% of them (n = 307) and courses lasting more than one day in 75.7% (n = 1,253). Surgical prophylaxis antimicrobials (n = 1,656) belonged most frequently to second-generation cephalosporins in 38.6% (n = 645) and were followed by glycopeptides in 11.4% (n = 190), imidazoles in 10.5% (n = 175), combinations of penicillins including beta-lactamase inhibitor in 9.7% (n = 162) and fluoroquinolones in 5.7% (n = 95). The antimicrobial agents that were used for surgical prophylaxis are presented in Fig. 3, ranked by frequency.

Frequent antimicrobial agents (ATC5 code) accounting for 75% of surgical prophylaxis indication

Of 9,003 antimicrobials recorded, 8,567 agents were antibacterials (95.2%) and 8,508 were classified in the AWaRe list (94.5% of total). Broad-spectrum antibacterials comprised 58.1% of antibacterials (4,979/8,567), and the ratio of broad-spectrum to narrow-spectrum antibiotics was 1.4. Prescribed agents came from the restricted classes of Watch in 63.4% (n = 5,397/8,508) and Reserve group in 13.3% (n = 1,127/8,508).

Among the 50 sample hospitals, the median HAI prevalence was 8.9% (95% CI: 5.5–10.5) with 25% and 75% percentiles of 3.5% and 12.4%, respectively. Descriptive statistics of independent and outcome variables assessed are presented in Supplementary Table 1 (ST1).

Results from univariable and multivariable linear regression analysis are presented in Supplementary Table 2 (ST2) and Table 3, respectively. After adjusting for covariates, HAI prevalence presented a statistically significant positive association with bed occupancy (β = 0.107, SE = 0.046, p = 0.027). Higher AU prevalence was associated with the secondary hospital type (β = 19.98, SE = 7.838, p = 0.014) in the multivariable analysis.