High rate of infections during ICU admission of patients with severe SARS-CoV-2 pneumonia: A matter of time?

Dear editor,

A recent meta-analysis published in this journal showed that the prevalence of bacterial co-infections in critically ill COVID-19 patients in Intensive Care Units (ICUs) hover around 14% (95% CI 5–26%). In contrast, other authors, applying molecular screening for co-infection, described co-infection rates as high as 28% and 41% in two ICUs. , In other meta-analysis, secondary bacterial infection during admission on hospitalized adult patients was identified in 14.3% of COVID-19 patients (95% CI 9.6–18.9%), although this prevalence might not reflect the rates in ICUs. In this work, we want to report our findings in a Spanish ICU with high incidence of both co-infection and secondary infections during the first wave of the pandemic.

We prospectively included all consecutive patients who were admitted to the ICU at the Araba University Hospital in Vitoria-Gasteiz (Spain) with the diagnosis of severe pneumonia caused by SARS-COV-2 between March 4th and June 2nd, 2020 (first wave). Outcomes of the first 48 patients were reported previoulsy. The study protocol was approved by the hospital Ethics and Clinical Research Committee, and informed consent was waived. All patients had a positive SARS-CoV-2 test using real-time reverse transcriptase PCR (Cobas Roche Diagnostics SLU) either from nasopharyngeal swabs or lower respiratory tract aspirates. Data were expressed as median and interquartile range or percentages as appropriate.

Ninety-two patients were admitted to the ICU during this period. Table 1 shows the demographic and clinical data of the patients. In 63 patients, broncho-alveolar lavage (BAL)/trach aspirates were collected for microbiologic culture, and in 33 of them (52%), an automated multiplex PCR test targeting 27 pathogens and 7 antimicrobial resistance genes was performed (analysis time of about 67 min; Film BIOFIRE® FILMARRAY® Respiratory 2.1 plus Panel, FA-RP). None of the 33 FA-RP tests (14 performed on admission) identified other respiratory viruses.

Table 1

Patient characteristics of infected and non infected patients.

Characteristic	Patients	No infection vs infection
Gender (male/female)	61/31 = 92
Age (mean ± SD)	64 ± 12
Days at ICU (median and IQR)	12 (5–25)	6 (2–11) vs 21 (12–37), p < 0.01
Days of ICU in patients with MV	17 (10–28)	9 (4–13) vs 22 (16–40), p < 0.01
Days of MV (median and IQR)	13 (8–26)	8 (4–11) vs 18 (12–18), p < 0.01
Scores II (mean ± SD):
APACHE	16 ± 5
SOFA	7 ± 3
Barthel	98 ± 7
Comorbidities (%)
Hypertension	40%
Obesity	40%
Smoking	37%
Lung disease	34%
Organ failure (%)
Hemodynamic	50%
Acute renal failure	41%
Multiple organ dysfunction syndrome	42%
Coagulopathy	24%
Liver failure	10%
Life support (%)
Vasopressors	72%	62% vs 83%, p < 0.05
Mechanical ventilation (MV)	75%	61% vs 85%, p < 0.01
Prone position	56%	42% vs 70%, p < 0.01
CRRT	8%
ECMO	2%
Death (%)	33%
Treatment (%)
Lopinavir/ritonavir	91%
Hydroxychloroquine	91%
Interferon	46%
Tocilizumab	26%
Antibiotics	90%	82% vs 96%, p < 0.03
Ceftriaxone	64%
Azithromycin	47%
Levofloxacin	37%
Piperacillin-tazobactam	32%	16% vs 43%, p < 0.01
Linezolid	27%	11% vs 39%, p < p < 0.04
Meropenem	20%	8% vs 28%, p < 0.03
Ceftazidime	11%	5% vs 33%, p < 0.01
Antifungal	12%	0% vs 20%, p < 0.01
Methylprednisolone	51%
Dexamethasone	17%

At admission or in the first 48 h of stay in the ICU, 32 microbial isolates were found in 24 patients (26%, 24/92). In these patients, concordant results between the FA-RP (≥ 104 DNA copies/ml) and cultures (BAL cut-off of 104 CFU/ml) were obtained in 11 of 14 patients (overall agreement =78%, kappa = 0.59 [95% CI 0.21–0.96]). Discordant results were obtained in 3 samples (Moraxella catarrhalis, Proteus spp and Streptococcus agalactiae). Table 2 shows the microbial isolates obtained from microbiologic cultures. The most frequently isolated microorganisms in respiratory samples were Staphylococcus aureus, Streptococcus pneumoniae and Haemophilus influenzae, microorganisms that commonly colonise the nasopharynx. Other identified isolates were Pseudomonas aeruginosa, Serratia marcescens and Enterococcus faecium (mostly in respiratory samples), microorganisms that are frequently nosocomial. These data contrast with the low rates reported by other authors in Spain , and focus on the peculiarity of critical patients also in this scenario. Likewise, it should be noted that the hospital stay before admission to the ICU of the patients with P. aeruginosa infection was markedly longer (median 9 days) than that of the general group (median 3 days).

Table 2

Clinical isolates from microbiologic cultures distributed by infectious site.

Isolates	RC		BC		UC		Catheter		Others		Total
S. pneumoniae	5	0	0	0	0	0	0	0	0	0	5	0
H. Influenzae	5	1	0	0	0	0	0	0	0	0	5	1
S aureus	5	6	1	0	1	0	0	0	0	0	7	6
CNS	0	2	0	5	0	9	0	3	0	0	0	19
E. faecalis	1	6	0	2	0	5	0	1	0	0	1	14
E. faecium	1	4	0	4	1	9	0	1	0	1	2	19
P. aeruginosa	4	14	0	3	0	3	0	1	0	1	4	22
M. catarrhalis	2	0	0	0	0	0	0	0	0	0	2	0
S. marcescens	3	1	1	0	0	0	0	0	0	0	4	1
Stenotrophomonas	1	5	1	0	0	0	0	0	0	0	2	5
E. coli	0	8	0	0	0	3	0	0	0	0	0	11
Candida species	0	7	0	1	0	6	0	0	0	1	0	15
Aspergillus	0	2	0	0	0	0	0	0	0	0	0	2
Klebsiella	0	2	0	1	0	0	0	0	0	1	0	4
Enterobacter spp	0	4	0	1	0	0	0	1	0	0	0	6
Total	27	62	3	17	2	35	0	7	0	4	32	125

RC: respiratory cultures (sputum or tracheal aspiration), BC: blood cultures, UC: urine cultures.

First column: Coinfection Second column; Secondary infection

On the other hand, 125 microbial isolates were found in 43 patients (47%, 43/92) during their stay in the ICU. Most samples were respiratory (52%), followed by urinary (22%), blood (18%) and catheter tips (8%). The most common isolated microorganisms were P. aeruginosa, E. faecium and Enterobacterales, which represent half of the isolates in all secondary infections. Candida spp isolated from respiratory samples and coagulase-negative staphylococci in blood cultures and urine cultures may be considered as normal microbiota or contaminants, respectively. The contaminant isolates may be explained by the use of personal protective equipment and the unfamiliarity of protocols by healthcare workers who do not usually work in the ICU. During admission, concordant results between the FA-RP and cultures were obtained in 12 out 19 patients (overall agreement  = 63%, kappa = 0.31 [95% CI -0.05–0.67]). Discordant results were obtained in 6 samples, E. faecalis (2), Aspergillus fumigatus (2), E. faecium (1) and Candida albicans (1), targets not included in the panel. Out the two patients with culture-positive of A. fumigatus, one was considered to be colonized, and the other one, which had a positive galactomannan serum antigen, was considered as possible case of COVID-19 associated pulmonary aspergillosis (CAPA), and treated accordingly. Even though the AspICU algorithm to diagnose invasive pulmonary aspergillosis in critically ill patients requires a positive respiratory culture to identify Aspergillus, the real prevalence remains elusive because of the current absence of a standardized definition for non-proven disease in non-neutropenic critically ill patients. , Nine multidrug-resistant strains were isolated, which represented 6% of microbial isolates: extended-spectrum beta-lactamase E. coli (4); multi-resistant P. aeruginosa (2), and methicillin-resistant S. aureus (3).

Regarding the consumption of antimicrobials, 90% of the patients received at least one antibiotic for a median of 6 days (IQR 2–10). Twelve percent of patients received antifungal treatment. Regarding the analytical data, the maximum levels of procalcitonin (Abbott Alinity I BRAHMS PCT) were significantly lower in patients who did not present infection (median 0.4 ng/mL (IQR 0.1–1.4) vs median 1.2 ng/mL (IQR 0.3–2.6). Although it is known that procalcitonin is a non-sensitive marker, it can be useful, framed in the appropriate clinical context, to suspect the presence of an associated infection.

Infected patients upon admission to the ICU presented significantly longer ICU stay (median 21 vs 6 days), required mechanical ventilation (83% vs 62%) and/or prone position (70% vs 40%) more frequently, and were the patients who received more antibiotics to treat nosocomial infections (see table 1). However, we cannot determine whether longer stays is a cause or a consequence of the development of infection. Finally, the overall mortality was 33%. without difference regardless the presence of an infection, in contrast to what it was observed by other authors. ,

In conclusion, based on our experience, the incidence of infections during SARS-Cov-2 infection, both at the beginning and during admission, is higher than that reported by other authors, which leads to greater morbidity, longer stay, higher antimicrobial use, and potential selection for resistant microorganisms. We must optimize the antimicrobial stewardship concept to tackle the challenge we face.