key: cord-0785905-7g1shsjp
authors: Budhiraja, Sandeep; Tarai, Bansidhar; Jain, Dinesh; Aggarwal, Mona; Indrayan, Abhaya; Das, Poonam; Mishra, RS; Bali, Supriya; Mahajan, Monica; Kirtani, Jay; Tickoo, Rommel; Soni, Pankaj; Nangia, Vivek; Lall, Ajay; Kishore, Nevin; Jain, Ashish; Singh, Omender; Singh, Namrita; Kumar, Ashok; Saxena, Prashant; Dewan, Arun; Aggarwal, Ritesh; Mehra, Mukesh; Jain, Meenakshi; Nakra, Vimal; Sharma, B D; Pandey, Praveen Kumar; Singh, YP; Arora, Vijay; Jain, Suchitra; Chhabra, Ranjana; Tuli, Preeti; Boobna, Vandana; Joshi, Alok; Aggarwal, Manoj; Gupta, Rajiv; Aneja, Pankaj; Dhall, Sanjay; Arora, Vineet; Chugh, Inder Mohan; Garg, Sandeep; Mittal, Vikas; Gupta, Ajay; Jyoti, Bikram; Sharma, Puneet; Bhasin, Pooja; Jain, Shakti; Singhal, RK; Bhasin, Atul; Vardani, Anil; Pal, Vivek; Pande, Deepak Gargi; Gulati, Tribhuvan; Nayar, Sandeep; Kalra, Sunny; Garg, Manish; Pande, Rajesh; Bag, Pradyut; Gupta, Arpit; Sharma, Jitin; Handoo, Anil; Burman, Purabi; Gupta, Ajay Kumar; Choudhary, Pankaj Nand; Gupta, Ashish; Gupta, Puneet; Joshi, Sharad; Tayal, Nitesh; Gupta, Manish; Khanna, Anita; Kishore, Sachin; Sahay, Shailesh; Dang, Rajiv; Mishra, Neelima; Sekhri, Sunil; Srivastava, Dr. Rajneesh Chandra; Agrawal, Dr. Mitali Bharat; Mathur, Mohit; Banwari, Akash; Khetarpal, Sumit; Pandove, Sachin; Bhasin, Deepak; Singh, Harpal; Midha, Devender; Bhutani, Anjali; Kaur, Manpreet; Singh, Amarjit; Sharma, Shalini; Singla, Komal; Gupta, Pooja; Sagar, Vinay; Dixit, Ambrish; Bajpai, Rashmi; Chachra, Vaibhav; Tyagi, Puneet; Saxena, Sanjay; Uniyal, Bhupesh; Belwal, Shantanu; Aier, Imliwati; Singhal, Mini; Khaduri, Ankit
title: Secondary infections modify the overall course of hospitalized COVID-19 patients: A retrospective study from a network of hospitals across North India
date: 2022-02-23
journal: IJID Regions
DOI: 10.1016/j.ijregi.2022.02.008
sha: f0e43683603b949ef2d2d6f3259d353c4d510feb
doc_id: 785905
cord_uid: 7g1shsjp

Objective : To get better insights into the extent of secondary bacterial and fungal infections in Indian hospitalized patients and to assess how these alter the course of COVID-19 so that the control measures can be suggested. Methods : This is a retrospective, multicentre study where data of all RT-PCR positive COVID-19 patients was accessed from Electronic Health Records (EHR) of a network of 10 hospitals across 5 North Indian states, admitted during the period from March 2020 to July 2021. Results : Of 19852 RT-PCR positive SARS-CO2 patients admitted during the study period, 1940 (9.8%) patients developed SIs. Patients with SIs were 8 years older on average (median age 62.6 years versus 54.3 years; P<0.001) than those without SIs. The risk of SIs was significantly (p < 0.001) associated with age, severity of disease at admission, diabetes, ICU admission, and ventilator use. The most common site of infection was urinary tract infection (UTI) (41.7%), followed by blood stream infection (BSI) (30.8%), sputum/BAL/ET fluid (24.8%), and the least was pus/wound discharge (2.6%). Gram negative bacilli (GNB) were the commonest organisms (63.2%), followed by Gram positive cocci (GPC) (19.6%) and fungus (17.3%). Most of the patients with SIs were on multiple antimicrobials – the most used were the BL-BLI for GNBs (76.9%) followed by carbapenems (57.7%), cephalosporins (53.9%) and antibiotics carbapenem resistant Enterobacteriaceae (47.1%). The usage of empirical antibiotics for GPCs was in 58.9% and of antifungals in 56.9% of cases, and substantially more than the results obtained by culture. The average stay in hospital for patients with SIs was almost twice than those without SIs (median 13 days versus 7 days). The overall mortality in the group with SIs (40.3%) was more than 8 times of that in those without SIs (4.6%). Only 1.2% of SI patients with mild COVID-19 at presentation died, while 17.5% of those with moderate disease and 58.5% of those with severe COVID-19 died (P< 0.001). The mortality was the highest in those with BSI (49.8%), closely followed by those with HAP (47.9%), and then UTI and SSTI (29.4% each). The mortality in diabetic patients with SIs was 45.2% while in non-diabetics it was 34.3% (p < 0.001). Conclusions : Secondary bacterial and fungal infections complicate the course of COVID-19 hospitalised patients. These patients tend to have a much longer stay in hospital, higher requirement for oxygen and ICU care, and significantly high mortality. The group most vulnerable to this complication are those with more severe COVID-19 illness, elderly, and diabetic patients. Judicious empiric use of combination antimicrobials in this set of vulnerable COVID-19 patients can save lives. It is desirable to have a region or country specific guidelines for appropriate use of antibiotics and antifungals to prevent their overuse.

1 Highlights:  Secondary infections can complicate the course of 10% of COVID-19 hospitalized patients.  Elderly, diabetic, and severe patients had the highest risk of secondary infections.  Hospital length of stay of patients with secondary infections was almost two times.  Patients with secondary infections had higher requirement of oxygen and ICU care.

The commonest site of secondary infection was UTI, followed by BSI.

The most common site of infection was urinary tract infection (UTI) (41.7%), followed by blood stream infection (BSI) (30.8%), sputum/BAL/ET fluid (24.8%), and the least was pus/wound discharge (2.6%). Gram negative bacilli (GNB) were the commonest organisms (63.2%), followed by Gram positive cocci (GPC) (19.6%) and fungus (17.3%). Most of the patients with SIs were on multiple antimicrobials -the most used were the BL-BLI for GNBs (76.9%) followed by carbapenems (57.7%), cephalosporins (53.9%) and antibiotics carbapenem resistant Enterobacteriaceae (47.1%). The usage of empirical antibiotics for GPCs was in 58.9% and of antifungals in 56.9% of cases, and substantially more than the results obtained by culture.

The average stay in hospital for patients with SIs was almost twice than those without SIs (median 13 days versus 7 days). The overall mortality in the group with SIs (40.3%) was more than 8 times of that in those without SIs (4.6%). Only 1.2% of SI patients with mild COVID-19 at presentation died, while 17.5% of those with moderate disease and 58.5% of those with severe COVID-19 died (P< 0.001). The mortality was the highest in those with BSI (49.8%), closely followed by those with HAP (47.9%), and then UTI and SSTI (29.4% each). The mortality in diabetic patients with SIs was 45.2% while in non-diabetics it was 34.3% (p < 0.001).

hospitalised patients. These patients tend to have a much longer stay in hospital, higher requirement for oxygen and ICU care, and significantly high mortality. The group most vulnerable to this complication are those with more severe COVID-19 illness, elderly, and diabetic patients. Judicious empiric use of combination antimicrobials in this set of vulnerable COVID-19 patients can save lives. It is desirable to have a region or country specific guidelines for appropriate use of antibiotics and antifungals to prevent their overuse.

Keywords: Secondary infections, bacteria, fungus, antimicrobials, Introduction:

Viral infections, particularly SARS-CoV-2, may predispose to concomitant and subsequent bacterial infections (Bengoechea and Bamford 2020, Manna et al. 2020) . Various explanations given for this phenomenon include direct damage to the respiratory epithelium caused by the virus, their effects on innate and adaptive immunity, and SARS-CoV-2 associated perturbation of gut homeostasis (Bengoechea and Bamford 2020, Manna et al. 2020) .

Secondary bacterial infections have been noted to be a significant contributor to increased morbidity and mortality in earlier influenza pandemics and during seasonal influenza, and also in other respiratory diseases (Morris et al. 2017 , Morens et al. 2008 . Shafran et al. (2021) found that COVID-19 patients had a higher rate of secondary bacterial infections compared to influenza patients (12.6% versus 8.7%, p=0.006). Other studies suggested that superinfections, especially in the later stage of illness, were encountered in 8% of patients with COVID-19, usually those that were more severely ill and those who died (Shafran et al. 2021 , Huang et al. 2020 , Chen et al. 2020 , Zhou et al. 2020 ).

Because of concerns of increased mortality in patients with bacterial superinfections during influenza pandemics, several guidelines advocate the use of empirical antibiotics for patients with severe COVID-19 (World Health Organization 2020, Alhazzani et al. 2020 ). This, however, has a potential of antibiotic overuse and increasing antimicrobial resistance (Huttner et al. 2020) . As the prevalence of secondary bacterial and fungal infections in COVID-19 patients in India is not clearly known, a better understanding would be crucial for treating COVID-19 and to help ensure responsible use of antimicrobials to minimize negative consequences of overuse.

The present study was undertaken to get better insights into the extent of secondary bacterial and fungal infections in Indian patients of COVID-19 admitted to hospitals and to assess how these alter the course of the disease. Evaluating the treatment strategies used in this group of patients may help design appropriate guidelines for empirical use of antimicrobials in COVID-19 Indian patients.

This is a retrospective, multicentre study where data of all RT-PCR positive COVID-19 patients was accessed from Electronic Health Records (EHR) of a network of 10 hospitals across 5 North Indian states, admitted during the period from March 2020 to July 2021. They were divided in to mild, moderate, and severe categories as per the government of India criteria (Ministry of Health and Family Welfare 2021). The data included demographic profile of patients, presence of diabetes, various investigations like CRP, D-Dimer, IL-6, ferritin, CPK, LDH, Trop-I and lymphocyte counts, wherever available, the HRCT chest severity score (CTSS), various treatment modalities like use of steroids, Remdesivir and convalescent plasma, average length of stay and in-hospital mortality.

Detailed data were available for secondary infections (SIs). The microbiological data in the form of culture results from blood, urine, pus/wound discharge, sputum, BAL fluid culture, and ET secretion cultures, was analysed and patients were categorized into four types of SIs, namely, blood stream infection (BSI), urinary tract infection (UTI), skin and soft tissue infection (SSTI) and hospital acquired pneumonia (HAP).

The patients with SIs were compared with those without SIs for the parameters just listed. Use of antibiotics, antifungals, and antivirals in those who developed secondary infections (SIs) was studied. Substantial number of patients was on multiple antimicrobials, and many had multiple sites of infections. For statistical analysis, these were included under one predominant or primary site of 6 infection if the same organism was isolated from different sites. Those who had more than two sites involved with same micro-organism were categorised to have primary site of infection, as clinically evident or supported by radiological evidence.

Various definitions that were used for the present study, are as follows: Hospital-acquired infection was defined as secondary infection occurring more than 48 h after hospitalization for SARS-CoV-2. Ventilator-associated pneumonia (VAP), a subset of HAP was defined as pneumonias occurring 48 hr of endotracheal intubation. For possible HAP/VAP diagnosis, indicators of worsening oxygenation (FiO2 value increase by ≥ 0.20 or PEEP value increase by ≥ 3 cm H2O) over 48 h and purulent respiratory secretions and/or a positive culture for a respiratory pathogen was required (Søgaard et al. 2021) .

Bloodstream infections (BSI) defined by the presence of viable bacterial or fungal microorganisms in the bloodstream (later demonstrated by the positivity of one or more blood cultures) that elicit or have elicited an inflammatory response characterized by the alteration of clinical, laboratory and hemodynamic parameters (Viscoli. 2016) .

Urinary tract infection (UTI) was defined as microbial infiltration of the otherwise sterile urinary tract and is one of the most common bacterial infections worldwide. UTIs encompass infections of the urethra (urethritis), bladder (cystitis), ureters (ureteritis), and kidney (pyelonephritis) (Barber et al. 2013) . We further segregated cases of isolated candiduria from the urine culture positive cases, which was defined as the prescence of candida >10 ^4 CFU/ml, together with pyuria (Gharaghani et al. 2018 ).

Skin and soft-tissue infections (SSTIs) encompass a variety of pathological conditions that involve the skin and underlying subcutaneous tissue, fascia, or muscle, ranging from simple superficial infections to severe necrotizing infections (Sartelli et al. 2018 ).

The data have been presented as counts and percentages for qualitative characteristics such as sex, place of admission, and use of oxygen, and as mean and SD for quantitative characteristics such as age. Length of stay in the hospital and laboratory parameters have been summarised in terms of median and inter-quartile rang (IQR) because of their highly skewed distribution. Statistical significance of the difference between cases with SI and without SI was assessed by chi-square test or Student t-test. Fisher exact test was used for comparing small (<5) frequencies. For highly skewed distributions, such as of inflammatory markers, Wilcoxon-Mann-Whitney test was used. A p-value less than 0.05 was considered statistically significant although, in this case, the number of cases is so large for some categories that p-values must be cautiously interpreted. SPSS 21 was used for calculations.

Demographics A total of 19852 RT-PCR positive SARS-COV2 patients were admitted during the period of the study in the network of our hospitals. Their records were retrieved from the electronic health record system. Of these, a total of 1940 (9.8%) patients developed secondary infections. No significant (p = 0.100) gender difference was observed but the patients with SIs were on average 8 years older than those without secondary infection (median age 62.6 years versus 54.3 years; p < 0.001) ( Table 1) . As the age increased, the incidence of SIs steeply increased from 4.0% in <45 years to 18.4% in ≥75 years. More than one-fifth (22.6%) of ICU patients were affected against only 3.0% ward patients (p < 0.001). Those on oxygen were significantly (p < 0.0010) more affected (12.1%) than those not on oxygen (4.9%) and the incidence increased with the increasing need of oxygen from O2 to noninvasive ventilation (NIV) to mechanical ventilation (MV). Such details for those on convalescent plasma therapy (CPT), steroids, and remdesivir are also given in Table 1 . The median length of stay was almost double (13 days) in the SI group compared to the group without SIs (7 days, p <0.001) ( Table 1) .

Of the 1940 patients with SIs, 598 (30.8%) had positive blood culture, 809 (41.7%) had positive urine culture, 51 (2.6%) had positive cultures from pus / wound discharge, and 482 (24.8%) had positive cultures from sputum / BAL fluid or ET secretions (Table2). Thus, the most common site was urine, followed by blood.

Out of 482 cases of HAP, 50 (10.4%) were on mechanical ventilator and 181 (37.6%) were on NIV. Out of 598 cases of BSI, 341 (57.0%) had central line inserted. Out of 809 cases of UTI, 197 (24.4%) had foley's catheter inserted.

Mean age of patients with infections at different sites was not significantly different (p = 0.315). Of all infections, HAP was significantly less in females (18.5%) compared to males (28.3%) whereas UTI was significantly more (48.9% vs. 37.8%, p < 0.0010). Those with diabetes had relatively more of BSI and UTI and less of SSTI and HAP. Table 3 . Candida albicans and Candida tropicalis were more common in urine and Candida Auris in blood.

Antibiotic sensitivity test was conducted among the samples tested for secondary infection. Carbapenem resistance was observed in 68% of Acinetobacter baumanii, 48% of Klebsiella pneumonia, 39% of E. coli, and 43% cases of Pseudomonas aeruginosa. Fluconazole resistance was studied in cultures positive for fungal infection and found fluconazole resistance in 54% of Candida Auris, 10% of Candida albicans, and 19% of non-albicans Candida.

The usage of various antimicrobial agents (antibiotics, antifungals, and antivirals) as initial empirical therapy in those patients who developed SIs was also studied. Almost all these patients were on multiple antibiotics and/or antifungals. In terms of usage of antimicrobials for COVID-19 infection, the commonly used medications were Remdesivir (74.8%), Favipiravir (21.2%), Doxycycline (50.2%), Ivermectin (43.5%) and Azithromycin (29.3%) . For empirical treatment of SIs, the most used antibiotics were those directed against GNBs. The most used antibiotics against GNBs were BL-BLIs (76.9%), carbapenems (57.7%), cephalosporins (53.9%), and antibiotics against CREs (47.1%).

Empirical usage of antibiotics against GPCs was seen in 58.9% of the patients with SIs. Interestingly, we observed empirical usage of antifungals in 56.9% of the patients with SIs. See Table 4 for details.

As shown in Table 5 , mortality (40.3%) in the group with SIs was more than 8 times the mortality (4.6%) in the group with no SIs. The proportion of patients getting SIs increased as the severity of COVID-19 increased and so did mortality. In mild COVID-19 group, only 166 patients (2.6%) had SIs, in moderate cases 628 (9.5%), and in severe cases 1146 (16.7%). Mortality in those with SIs with mild, moderate, and severe COVID-19 showed steep increase at 1.2%, 17.5% and 58.5%, respectively ( Table 5 ).

The mortality in relation to the site of infection showed highest mortality (49.8%) in patients with blood stream infection, closely followed by 47.9% with pneumonia, and 29.4% each with urinary and skin / soft tissue infections (Table 6 ).

There were 241 (82.6%) deaths from amongst 341 cases of BSI which had central line inserted vs a mortality rate of 17.4% in cases of BSI without central line.

The mortality in patients with only one identified organism was 37.8% against 56.3% in patients with more than one organism and was significantly associated with the site of infection in both the cases (p < 0.001). The proportionate pattern of mortality in cases with single and multiple organisms was nearly similar for all sites of infection although the differences were statistically significant (p < 0.001) because of large sample in our study (Table 6 ). Similarly, mortality in patients with one site of infection was 28.8% against 62.5% in patients with multiple sites of infections (p < 0.001).

More than half (1066, 54.9%) of 1940 who developed SIs were diabetics. The mortality in them was 45.2% against 34.3% in those without diabetes (Table 7) and this difference was statistically significant (p < 0.001).

The trend of various inflammatory markers commonly used for monitoring COVID-19 admitted patients showed a higher value in those with SIs versus those without. The median values (Table 8) for CRP, D-dimer, ferritin, IL-6, LDH, and CPK were higher and ALC lower in the group with infection. This difference was even more if we compare the values for those patients who died against those who survived, across both the groups. However, the difference in the median levels of inflammatory markers in those who died in both the groups was not very different and, in some markers (such as CRP, IL-6, LDH and CPK), the median values were actually higher in those who had no secondary infection and died. The median CTSS for the overall group with SIs was 15 (IQR: 10-19) and that for the group without SIs was 10 (IQR: 7-14). Again, in those who died in both the groups, the CTSS score was almost similar (median 17) (Table 8 ).

COVID-19 patients are at a higher risk of secondary bacterial and fungal infections, and these are associated with increased morbidity and mortality (Bengoechea and Bamford 2020 , Manna et al. 2020 , Morris et al. 2017 , Morens et al. 2008 . Several factors are known to contribute to higher risk of secondary infections in these patients. The damage to the respiratory epithelium caused by the virus, as well as their effects on innate and adaptive immunity, antagonising IFN responses that enhance bacterial adherence, colonisation, growth, and invasion into healthy sites in the respiratory tract, are important mechanisms (Bengoechea and Bamford 2020, Manna et al. 2020) . Down regulation and differential regulation of immune genes are mechanisms that may create a conducive environment for occurrence of secondary bacterial infections, favouring bacterial attachment to host structural cells and pro-inflammatory environment conducive to suppression of anti-bacterial host defences. In addition, Bengoechea and Bamford (2020) suggested SARS-CoV-2-associated perturbation of gut homeostasis as a mechanism that may potentially affect the disease outcomes in patients with severe COVID-19 infection, including predisposing to secondary lung infections.

Overall, 9.8% of the total hospitalised COVID-19 patients in our network were diagnosed with secondary bacterial or fungal infections. A retrospective study by Vijay et al. (2021) on 17,534 COVID-19 patients admitted in 10 hospitals of ICMR-AMR Surveillance network in India, reported SIs in only 3.6% cases. A meta-analysis of 24 studies, including 3338 patients with COVID-19, done by Langford et al. (2020) reported overall bacterial infection in 6.9%, and 8.1 % in critically ill patients. Shafran et al. (2021) studied 1384 cases (642 COVID-19 cases and 742 influenza cases) for blood and sputum culture results, clinical parameters and outcomes and compared these parameters between the COVID-19 cases and influenza cases. Higher rate of bacterial infection was found in COVID-19 than in those infected with Influenza (12.6% vs. 8.7%). A review of secondary pulmonary infection in patients with COVID-19 pneumonia by Chong et al. (2021) from USA reported the incidence of secondary pulmonary infection to be 16% for bacterial infection while 6.3% for fungal infection. Secondary pulmonary infection was predominantly in critically ill hospitalized cases in their study. Thus, the incidence of SIs in COVID-19 patients could be in the range of 5% to 15%, and it increased with the severity of disease.

We could identify several factors that increased the risk of developing SIs in COVID-19 patients. Average age of patients who had SIs was 8 years more than those without SIs (median 62.6 years versus 54.3 years) (p < 0.001). Also, as age increased, so did the risk of getting SIs. Only 4% of those under 45 years got SIs whereas 18.4% patients above 75 years of age developed SIs. In the earlier Indian study by Vijay et al. (2021) the mean age of COVID patients diagnosed with SI was 53.3 ± 9.36 yrs. We observed a higher age of the SI patients and a clear upward gradient of SI incidence with increasing age.

In our present study, 54.9% patients were diabetic in the SIs group whereas the overall prevalence of diabetes in COVID-19 these patients was 43.8% (Budhiraja et al. 2021) . This difference is statistically significant (p < 0.001). A study from USA by Adelman et al. (2021) reported for their cohort of 774 COVID-19 cases that hypertension was in 75.5% and diabetes mellitus in 45.7% cases. A case-control study conducted in Pakistan by Nasir et al. (2021) also reported diabetes mellitus and hypertension as the most common comorbidities in SI cases. In our study, we found that patients with diabetes had relatively more of BSI and UTI and less of SSTI and HAP.

We found a clear correlation between the severity of COVID-19 at the time of admission and the risk of getting SIs. Only 2.6% of those with mild disease developed SIs, while moderate and severe disease had SIs 9.5% and 16.7%, respectively. This may be partly related to the need for hospitalisation and ICU stay in the cases with moderate and severe disease. Nasir et al. (2021) reported that the critically ill cases at the time of admission were at 4.42 times higher risk for bacterial infection. Chong et al. (2021) also found secondary pulmonary infections predominantly in critically ill hospitalized cases. There is a clear relation of higher incidence of SIs with the severity of COVID-19 at admission.

In our study, 22.6% of patients admitted to ICU developed SIs, while only 3% of those admitted in ward got SIs (p < 0.001). Vijay et al. (2021) reported that among the cases with confirmed SIs, 71.7% were in ICU while 28.3% were in ward at the time of diagnosis of SI. ICU admission seems to have a definite association with SIs.

It may be difficult to draw a cause-effect relation between the need for oxygen and the risk of getting SIs, but we observed that SIs developed in 12.1% of patients on oxygen and only 4.9% of those not on oxygen got SIs (P< 0.001). Similarly, the risk of getting SIs increased with increasing need of oxygen and ventilator support. The risk of getting SIs for patients on oxygen by nasal prongs/face mask was 8.4%, for those on NIV was 21.7%, and for those on MV was 23%. This could also reflect more severe disease in those COVID-19 patients who got serious SIs and hence needed to be on a greater support.

The most common site of SI in our study was urine (41.7%), followed by blood (41.7%), and pneumonia (24.8%). Skin infection was the least common (2.6%). This spectrum is different from the one reported by Vijay et al. (2021) with blood and respiratory as the most common sites of SIs.

Almost 13.4% of the patients had infection with more than one microorganism and 34.1% had multiple sites of infection. Shafran et.al. (2021) reported presence of more than 1 coinfection in only 4.5% of SARS-CoV-2 cases. Adelman et al. (2021) reported that 30.7% cases required mechanical ventilation and out of these 27.3% had positive respiratory culture, with Staphylococcus aureus (34.5%) being the most common bacteria followed by Pseudomonas aeruginosa (19.0%) and Klebsiella spp. (16.7%). Out of 774 cases, blood sample culture was positive in 76% cases and 4.7% (36) had blood stream infection -majority being in ICU (66.7%; 24/36 cases). Shafran et al. (2021) reported 85% of isolates to be positive in blood culture while 14.2 % were in respiratory sample. Vijay et al. (2021) , from India, found overall Gram-negative pathogens (78.03%) as the most predominant isolated pathogen, in which most common isolates were Klebsiella pneumoniae (29.3%) followed by Acinetobacter baumanii (21.07%), Pseudomonas aeruginosa (9.6%) and E. coli (8.2%). Candida spp. were isolated from 6% of admitted cases, of which 1.2% were of Candida Auris. Klebsiella pneumoniae was the most common isolate in blood (29.7%) and respiratory specimen (35%), while in urine most common isolate was E. coli (27.17%) followed by Klebsiella pneumonia (18.4%) and Candida spp. (18.4%). Such varying findings indicate that the spectrum of SIs may be population specific.

We studied the number of patients developing SIs in relation to various treatment modalities used for treatment of COVID-19 patients. This may not necessarily mean that these drugs increased the risk of SIs but may simply reflect high usage of these medicines, especially in sicker patients. Among those patients who received steroids, 10.3% had SIs, while only 3.1% in the group that did not receive steroids. More than one-fourth (26.9%) of those who received convalescent plasma (CP) got SIs, while only 7.3% had SIs in the group that did not receive CP. In the group that got remdesivir, 10.6% had SIs while 8.5% of those who did not receive had SIs. Nasir et al. (2021) noted use of systemic steroids to be in significantly higher proportion of cases with bacterial infection than in those without bacterial infection (92% vs. 62%).

Most of our patients with SIs were on multiple antimicrobials. The most used antibiotics were against GNBs. BL-BLI combination therapy was found to be the commonest used treatment (76.9%), followed by carbapenems (57.7%) and cephalosporins (53.9%). Antibiotics directed against GNB-CRE organisms such as Polymyxin B, Colistin, Fosfomycin, Minocycline and Tigecycline were used in 47.1% patients. This matched with the microbiological flora as we identified GNBs to be the cause of SIs in 63.2% cases. The high prevalence of CRE in GNBs (68% in Acinetobacter baumanii, 48% in Klebsiella peumoniae, 39% in E. coli and 43% in Pseudomonas aeruginosa) justified the empirical usage of antibiotics against CREs. The usage of antibiotics against GPCs (Staphylococcus, CONS, Enterococcus) in our study was found high at 58.9%, whereas the actual microbiological culture data revealed these organisms could be identified in only 19.6% samples. Similarly, antifungals were used in 56.9% cases in the present study, while the fungus (Candida sp) was isolated in only 17.3% cases. High degree of azole resistance in various Candida species (54% fluconazole resistance in Candia auris,19% resistance in non-albicans Candida and 10% in Candida albicans), and high level of isolation of Candida Auris (25.7% of Candida isolates in BSI and 13.4% in UTI), Candida tropicalis (30.1% in BSI and 45.1% in UTI), justifies the empirical use of Echinocandins (10.3%) and . However, there is a scope of significant improvement in terms of rationalizing the usage of antimicrobials, especially empiric coverage against GPC and fungus. Each country needs to develop their empiric antibiotic guidelines for hospitalised COVID-19 patients for optimizing the therapy and reduce the potential harm caused by future development of antimicrobial resistance.

Drug resistance profile of the isolated pathogen was studied by Vijay et al. (2021) and found 47.1% were infected with multiple drug resistant organisms. -74.2% of GNB isolates were resistant to carbapenems alone. They reported the use of third generation cephalosporins (16.6%), β-lactam-βlactamase inhibitors combination (57.3%), and carbapenems (43.7%) in the management of COVID-19 with SI. Vancomycin or teicoplanin was prescribed to 24.9% patients. They also reported that the empirical cover for Gram-positive pathogens may not be warranted as the SIs were predominantly caused by Gram-negative pathogens (78.3%) in their cohort. They also found that 10% patients received antifungals without any evidence of fungal infection. Shafran et al. (2021) found that culture reports in cases with either influenza or COVID-19 showed Pseudomonas aeruginosa and Staphylococcus aureus as the most common secondary bacterial infections. Gram negative represented 75% of in both groups. Interestingly, enterococcus infection was found to be more prevalent in cases of COVID-19 than in influenza cases (8.6% vs. 0%), and late infection with gram positive bacteria was also more common in cases with COVID-19 infection. Langford et al. (2020) analysed the use of antibiotics in COVID-19 patients and found that over 70% cases received antibiotics, with majority constituted by broad spectrum antibiotics like third generation cephalosporins and fluoroquinolones. Adelman et al. (2021) reported that the most common organism isolated were gram negative (28.6%), Staphylococcus aureus (16.7%), Candida species (16.7%) and CONS (11.9%). Nearly 50% were central line associated BSI (CLBSI). Nasir et al. (2021) found that among the bacterial infection, gram negative (85%) were more common than gram positive organism. Most frequent organism isolated from blood was MDR Acinetobacter followed by E. coli, Enterococcus, and Klebsiella pneumonia. Among cases with secondary bacterial HAP, the most common isolate was MDR Acinetobacter followed by MDR Pseudomonas aeruginosa. Chong et al. reported that use of antibiotics was in 60-100% cases in the studies they reviewed, and the most common bacterial microorganism was Pseudomonas aeruginosa (21%), followed by Klebsiella species (17.2%), Staphylococcus aureus (13.5%), E. coli (10%), and Stenotrophomonas maltophilia (3%). Aspergillus fumigates was the most frequently isolated species among fungal infection in COVID-19 cases. Most of the studies showed MDR-GNs to be the commonest organism causing SIs in COVID-19 patients.

The average length of hospitalisation in our SI cases was twice as high as in those without SIs (13 days versus 7 days; p < 0.001). We could identify several risk factors which increased mortality by more than 8 times in SI cases (40.3% vs. 4 .6% in the group without SIs). Vijay et al. (2021) noted that among cases of COVID-19 with SI, mortality was higher in critically ill patients (68%) compared to the patients in wards (27.6%). Nasir et al. (2021) found that cases with COVID-19 having bacterial infection had comparatively greater proportion of deaths compared to controls (42% vs. 18%). The SIs are significant factor for mortality, and they are mostly treatable. Thus, at least some of these deaths can be avoided.

Among the patients who had SIs, severity of COVID-19 disease at the time of admission had correlation with mortality. The mortality in mild, moderate, and severe disease was 1.2%, 17.5% and 58.5%, respectively. The mortality anyways is expected to rise with severity of COVID-19, but SIs may have contributed to the steep rise in the gradient. The mortality in the patients with SIs who had diabetes was 45.2% while in the group without diabetes was 34.3% (p < 0.001). The mortality in the group with BSI was highest (49.8%), followed closely in HAP (47.9%) and was 29.4% each in SSTI and UTI group. Adelman et al. found significantly higher overall mortality in COVID-19 cases with BSI compared to those without any BSI (50% vs. 13.8%). However, they did not find any significant difference in mortality rates among the intubated cases with or without identified bacterial respiratory pathogen.

Mortality in our group of patients with SIs, who had only one identifiable microorganism, was 37.8%, which climbed to 56.3% in patients with more than one microorganism (p < 0.001). Shafran et al. (2021) reported an overall mortality (in both COVID-19 and influenza cases combined) of 13.2 % in cases without infection, while mortality was 33% and 61% in cases with one infection and in cases with two infections, respectively. They however, found that in COVID-19 group, mortality was 48.1% in cases with one infection and 75.9% in cases with more than one infection. Patients with SIs at only one site in our series had a mortality of 28.8%, which rose to 62.5% in those with multiple sites of infection (p < 0.001). These finding suggest that secondary infection are significant contributing factor for disease severity among COVID-19 patients leading to higher mortality.

The median values for CRP, D-dimer, ferritin, IL-6, LDH and CPK were higher and ALC lower in the group with infection in our study. This difference was even more if we compare the values for those patients who died against those who survived, across both groups. However, the difference in the median levels of inflammatory markers in those who died in both the groups was not very different and for some markers (such as CRP, IL-6, LDH and CPK) the median values were actually higher in those who had no secondary infection and died. Cytokine storm causing significant elevation of these inflammatory markers, independent of secondary infections, would be the most likely reason for this. Nasir et al. (2021) found median C-reactive protein (169 vs. 81) and median NLR (8 vs. 4) to be significantly higher in cases with SIs while no significant difference was noted in procalcitonin level (0.36 vs. 0.14) in COVID-19 cases with bacterial infection compared to those without bacterial infection. The median CTSS for the overall group with SIs was 15 (IQR: 10-19) and that for the group without SIs was 10 (IQR: 7-14). Again, in those who died in both the groups, the CTSS score was almost similar (median 17).

hospitalised patients. These patients tend to have a much longer stay in hospital, higher requirement for oxygen and ICU care, and significantly high mortality. The group most vulnerable to this complication are those with more severe COVID-19 illness, elderly, and diabetic patients. Judicious empiric use of combination antimicrobials in this set of vulnerable COVID-19 patients can save lives. It is desirable to have a region or country specific guidelines for appropriate use of antibiotics and antifungals to prevent their overuse. 

Secondary bacterial pneumonias and bloodstream infections in patients hospitalized with COVID-19

Surviving sepsis campaign: guidelines on the management of critically ill adults with coronavirus disease 2019 (COVID-19)

Urinary Tract Infections: Current and Emerging Management Strategies

SARS-CoV-2, bacterial co-infections, and AMR: The deadly trio in COVID-19?

Differentials in the characteristics of COVID-19 cases in Wave-1 and Wave-2 admitted to a network of hospitals in North India

State-of-the-art review of secondary pulmonary infections in patients with COVID-19 pneumonia

Candiduria; a review article with specific data from Iran

Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. The Lancet

COVID-19: Don't neglect antimicrobial staweardship principles!

Bacterial co-infection and secondary infection in patients with COVID-19: a living rapid review and meta-analysis

Molecular pathogenesis of secondary bacterial infection associated to viral infections including SARS-CoV-2

Clinical Management Protocol: COVID-19

Predominant role of bacterial pneumonia as a cause of death in pandemic influenza: Implications for pandemic influenza preparedness

Secondary bacterial infections associated with influenza pandemics

Risk factors for bacterial infections in patients with moderate to severe COVID-19: A case-control study

WSES/SIS-E consensus conference: recommendations for the management of skin and soft-tissue infections

Secondary bacterial infection in COVID-19 patients is a stronger predictor for death compared to influenza patients

Community-acquired and hospital-acquired respiratory tract infection and bloodstream infection in patients hospitalized with COVID-19 pneumonia

Secondary infections in hospitalized COVID-19 patients: Indian experience. Infection and Drug Resistance

Bloodstream infections: The peak of the iceberg

Clinical management of COVID-19: Interim guidance

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Ceftriaxone-Sulbactum (373, 25%)

44.8%), Cefepime (326, 31.2%), Cefuroxime (170, 16.3%), Others (81, 7.8%)

Vancomycin (38, 3.4%), Linezolid (344, 30.0%)

Antibiotics for Anaerobes Clindamycin (138, 45.1%), Metronidazole (168, 54.9%)

Gentamycin (22, 15.3%)

31.1%), Ofloxacin (10, 22.2%)

Other Antibiotics for GNB Trimethoprim-Sulfamethoxazole (50, 55.6%), Nitrofurantoin (38, 42.2%)

Doxycyline (973, 50.2%), Ivermectin (843, 43.5%)

Fluconazole (410, 61.4%)

Antivirals Remdesivir (1273, 74.8%), Favipiravir (360, 21.2%), Oseltamivir (40, 2.3%), Lopinavir-Ritonavir (11, 0.6%)

Note: Most of the patients in SI group were on multiple antibiotics and/or antifungals CRP -n