Machine Learning Highlights Downtrending of COVID-19 Patients with a Distinct Laboratory Profile

Background New York City (NYC) experienced an initial surge and gradual decline in the number of SARS-CoV-2-confirmed cases in 2020. A change in the pattern of laboratory test results in COVID-19 patients over this time has not been reported or correlated with patient outcome. Methods We performed a retrospective study of routine laboratory and SARS-CoV-2 RT-PCR test results from 5,785 patients evaluated in a NYC hospital emergency department from March to June employing machine learning analysis. Results A COVID-19 high-risk laboratory test result profile (COVID19-HRP), consisting of 21 routine blood tests, was identified to characterize the SARS-CoV-2 patients. Approximately half of the SARS-CoV-2 positive patients had the distinct COVID19-HRP that separated them from SARS-CoV-2 negative patients. SARS-CoV-2 patients with the COVID19-HRP had higher SARS-CoV-2 viral loads, determined by cycle threshold values from the RT-PCR, and poorer clinical outcome compared to other positive patients without the COVID12-HRP. Furthermore, the percentage of SARS-CoV-2 patients with the COVID19-HRP has significantly decreased from March/April to May/June. Notably, viral load in the SARS-CoV-2 patients declined, and their laboratory profile became less distinguishable from SARS-CoV-2 negative patients in the later phase. Conclusions Our longitudinal analysis illustrates the temporal change of laboratory test result profile in SARS-CoV-2 patients and the COVID-19 evolvement in a US epicenter. This analysis could become an important tool in COVID-19 population disease severity tracking and prediction. In addition, this analysis may play an important role in prioritizing high-risk patients, assisting in patient triaging and optimizing the usage of resources.

Comparison of a Target Trial Emulation Framework vs Cox Regression to Estimate the Association of Corticosteroids With COVID-19 Mortality.

Importance: Communication and adoption of modern study design and analytical techniques is of high importance for the improvement of clinical research from observational data. Objective: To compare a modern method for statistical inference, including a target trial emulation framework and doubly robust estimation, with approaches common in the clinical literature, such as Cox proportional hazards models. Design, Setting, and Participants: This retrospective cohort study used longitudinal electronic health record data for outcomes at 28-days from time of hospitalization within a multicenter New York, New York, hospital system. Participants included adult patients hospitalized between March 1 and May 15, 2020, with COVID-19 and not receiving corticosteroids for chronic use. Data were analyzed from October 2021 to March 2022. Exposures: Corticosteroid exposure was defined as more than 0.5 mg/kg methylprednisolone equivalent in a 24-hour period. For target trial emulation, exposures were corticosteroids for 6 days if and when a patient met criteria for severe hypoxia vs no corticosteroids. For approaches common in clinical literature, treatment definitions used for variables in Cox regression models varied by study design (no time frame, 1 day, and 5 days from time of severe hypoxia). Main Outcomes and Measures: The main outcome was 28-day mortality from time of hospitalization. The association of corticosteroids with mortality for patients with moderate to severe COVID-19 was assessed using the World Health Organization (WHO) meta-analysis of corticosteroid randomized clinical trials as a benchmark. Results: A total of 3298 patients (median [IQR] age, 65 [53-77] years; 1970 [60%] men) were assessed, including 423 patients who received corticosteroids at any point during hospitalization and 699 patients who died within 28 days of hospitalization. Target trial emulation analysis found corticosteroids were associated with a reduced 28-day mortality rate, from 32.2%; (95% CI, 30.9%-33.5%) to 25.7% (95% CI, 24.5%-26.9%). This estimate is qualitatively identical to the WHO meta-analysis odds ratio of 0.66 (95% CI, 0.53-0.82). Hazard ratios using methods comparable with current corticosteroid research range in size and direction, from 0.50 (95% CI, 0.41-0.62) to 1.08 (95% CI, 0.80-1.47). Conclusions and Relevance: These findings suggest that clinical research based on observational data can be used to estimate findings similar to those from randomized clinical trials; however, the correctness of these estimates requires designing the study and analyzing the data based on principles that are different from the current standard in clinical research.

A metagenomic DNA sequencing assay that is robust against environmental DNA contamination.

Metagenomic DNA sequencing is a powerful tool to characterize microbial communities but is sensitive to environmental DNA contamination, in particular when applied to samples with low microbial biomass. Here, we present Sample-Intrinsic microbial DNA Found by Tagging and sequencing (SIFT-seq) a metagenomic sequencing assay that is robust against environmental DNA contamination introduced during sample preparation. The core idea of SIFT-seq is to tag the DNA in the sample prior to DNA isolation and library preparation with a label that can be recorded by DNA sequencing. Any contaminating DNA that is introduced in the sample after tagging can then be bioinformatically identified and removed. We applied SIFT-seq to screen for infections from microorganisms with low burden in blood and urine, to identify COVID-19 co-infection, to characterize the urinary microbiome, and to identify microbial DNA signatures of sepsis and inflammatory bowel disease in blood.

Predictors of Covid-19 Vaccination Response After In-Vivo T-Cell-Depleted Stem Cell Transplantation.

Covid-19 vaccination is recommended in allogeneic transplant recipients, but many questions remain regarding its efficacy. Here we studied serologic responses in 145 patients who had undergone allogeneic transplantation using in vivo T-cell depletion. Median age was 57 (range 21-79) at transplantation and 61 (range 24-80) at vaccination. Sixty-nine percent were Caucasian. One third each received transplants from HLA-identical related (MRD), adult unrelated (MUD), or haploidentical-cord blood donors. Graft-versus-host disease (GVHD) prophylaxis involved in-vivo T-cell depletion using alemtuzumab for MRD or MUD transplants and anti-thymocyte globulin for haplo-cord transplants. Patients were vaccinated between January 2021 and January 2022, an average of 31 months (range 3-111 months) after transplantation. Sixty-one percent received the BNT162b2 (bioNtech/Pfizer) vaccine, 34% received mRNA-1273 (Moderna), and 5% received JNJ-78436735 (Johnson & Johnson). After the initial vaccinations (2 doses for BNT162b2 and mRNA-1273, 1 dose for JNJ-7843673), 124 of the 145 (85%) patients had a detectable SARS-CoV-2 spike protein (S) antibody, and 21 (15%) did not respond. Ninety-nine (68%) had high-level responses (≥100 binding antibody units [BAU]/mL)m and 25 (17%) had a low-level response (<100 BAU/mL). In multivariable analysis, lymphocyte count less than 1 × 109/ mL, having chronic GVHD, and being vaccinated in the first year after transplantation emerged as independent predictors for poor response. Neither donor source nor prior exposure to rituximab was predictive of antibody response. SARS-CoV-2 vaccination induced generally high response rates in recipients of allogeneic transplants including recipients of umbilical cord blood transplants and after in-vivo T cell depletion. Responses are less robust in those vaccinated in the first year after transplantation, those with low lymphocyte counts, and those with chronic GVHD.

Impact of Severe Acute Respiratory Syndrome Coronavirus 2 Viral Load on Risk of Intubation and Mortality Among Hospitalized Patients With Coronavirus Disease 2019.

BACKGROUND: Patients hospitalized with coronavirus disease 2019 (COVID-19) frequently require mechanical ventilation and have high mortality rates. However, the impact of viral burden on these outcomes is unknown. METHODS: We conducted a retrospective cohort study of patients hospitalized with COVID-19 from 30 March 2020 to 30 April 2020 at 2 hospitals in New York City. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) viral load was assessed using cycle threshold (Ct) values from a reverse transcription-polymerase chain reaction assay applied to nasopharyngeal swab samples. We compared characteristics and outcomes of patients with high, medium, and low admission viral loads and assessed whether viral load was independently associated with intubation and in-hospital mortality. RESULTS: We evaluated 678 patients with COVID-19. Higher viral load was associated with increased age, comorbidities, smoking status, and recent chemotherapy. In-hospital mortality was 35.0% (Ct <25; n = 220), 17.6% (Ct 25-30; n = 216), and 6.2% (Ct >30; n = 242) with high, medium, and low viral loads, respectively (P < .001). The risk of intubation was also higher in patients with a high viral load (29.1%) compared with those with a medium (20.8%) or low viral load (14.9%; P < .001). High viral load was independently associated with mortality (adjusted odds ratio [aOR], 6.05; 95% confidence interval [CI], 2.92-12.52) and intubation (aOR, 2.73; 95% CI, 1.68-4.44). CONCLUSIONS: Admission SARS-CoV-2 viral load among hospitalized patients with COVID-19 independently correlates with the risk of intubation and in-hospital mortality. Providing this information to clinicians could potentially be used to guide patient care.

Changes in SARS-CoV-2 viral load and mortality during the initial wave of the pandemic in New York City.

Public health interventions such as social distancing and mask wearing decrease the incidence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, but it is unclear whether they decrease the viral load of infected patients and whether changes in viral load impact mortality from coronavirus disease 2019 (COVID-19). We evaluated 6923 patients with COVID-19 at six New York City hospitals from March 15-May 14, 2020, corresponding with the implementation of public health interventions in March. We assessed changes in cycle threshold (CT) values from reverse transcription-polymerase chain reaction tests and in-hospital mortality and modeled the impact of viral load on mortality. Mean CT values increased between March and May, with the proportion of patients with high viral load decreasing from 47.7% to 7.8%. In-hospital mortality increased from 14.9% in March to 28.4% in early April, and then decreased to 8.7% by May. Patients with high viral loads had increased mortality compared to those with low viral loads (adjusted odds ratio 2.34). If viral load had not declined, an estimated 69 additional deaths would have occurred (5.8% higher mortality). SARS-CoV-2 viral load steadily declined among hospitalized patients in the setting of public health interventions, and this correlated with decreases in mortality.

Sex-Related Differences in Clinical Presentation and Risk Factors for Mortality in Patients Hospitalized With Coronavirus Disease 2019 in New York City.

We evaluated sex-related differences in symptoms and risk factors for mortality in 4798 patients hospitalized with coronavirus disease 2019 in New York City. When adjusted for age and comorbidities, being male was an independent predictor of death with mortality significantly higher than females, even with low severe acute respiratory syndrome coronavirus 2 viral load at admission.

Bacterial Coinfections in Coronavirus Disease 2019.

Bacterial coinfections increase the severity of respiratory viral infections and were frequent causes of mortality in influenza pandemics but have not been well characterized in patients with coronavirus disease 2019 (COVID-19). The aim of this review was to identify the frequency and microbial etiologies of bacterial coinfections that are present upon admission to the hospital and that occur during hospitalization for COVID-19. We found that bacterial coinfections were present in <4% of patients upon admission and the yield of routine diagnostic tests for pneumonia was low. When bacterial coinfections did occur, Staphylococcus aureus, Streptococcus pneumoniae, and Haemophilus influenzae were the most common pathogens and atypical bacteria were rare. Although uncommon upon admission, bacterial infections frequently occurred in patients with prolonged hospitalization, and Pseudomonas aeruginosa, Klebsiella spp., and S. aureus were common pathogens. Antibacterial therapy and diagnostic testing for bacterial infections are unnecessary upon admission in most patients hospitalized with COVID-19, but clinicians should be vigilant for nosocomial bacterial infections.

Failure of chronic hydroxychloroquine in preventing severe complications of COVID-19 in patients with rheumatic diseases.

OBJECTIVE: To compare baseline characteristics, clinical presentations and outcomes of patients with rheumatic conditions requiring hospitalization for coronavirus disease 2019 (COVID-19) who received chronic HCQ with those who did not receive chronic HCQ. METHODS: We identified all patients with a rheumatologic disease who were admitted with COVID-19 to two hospitals in New York City between 3 March 3 and 30 April 2020. Patients who received chronic HCQ prior to admission were matched 1:2 (±10 years of age) with patients who did not receive chronic HCQ. We compared demographics, comorbidities, HCQ dosages, concurrent medications, presentations and outcomes between the groups. RESULTS: There were 14 patients receiving HCQ and 28 matched control subjects. The median age of cases was 63 years [interquartile range (IQR) 43-73) and 60 years (IQR 41-75) for controls. Control subjects had a higher prevalence of pulmonary diseases (42.8%), diabetes (35.7%) and obesity (35.7%) than their case counterparts (28.6%, 14.3% and 7.1%, respectively). A higher proportion of cases than control subjects (50% vs 25%) reported the use of prednisone for their rheumatic conditions prior to admission. Despite these differences in baseline characteristics, univariate logistic regression revealed no statistically significant differences in the need for mechanical ventilation [OR 1.5 (95% CI 0.34, 6.38)] or in-hospital mortality [OR 0.77 (95% CI 0.13, 4.56)]. CONCLUSION: HCQ therapy in individuals with rheumatic conditions was not associated with less severe presentations of COVID-19 among hospitalized patients compared with individuals with rheumatic conditions not receiving HCQ.

Comparison of Two High-Throughput Reverse Transcription-PCR Systems for the Detection of Severe Acute Respiratory Syndrome Coronavirus 2.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has emerged as the cause of a worldwide pandemic. Many commercial SARS-CoV-2 reverse transcription-PCR (RT-PCR) assays have received Emergency Use Authorization from the U.S. Food and Drug Administration. However, there are limited data describing their performance, in particular the performance of high-throughput SARS-CoV-2 RT-PCR systems. We analyzed the diagnostic performance of two high-throughput systems: cobas 6800 and Panther Fusion, and their associated RT-PCR assays, with a collection of 389 nasopharyngeal specimens. The overall agreement between the platforms was 96.4% (375/389). Cohen's kappa analysis rated the strength of agreement between the two platforms as "almost perfect" (κ = 0.922; standard error, 0.051). Furthermore, there was no significant difference between corresponding cycle threshold values generated on the two systems (P value = 0.88; Student's t test). Taken together, these data imply that the two platforms can be considered comparable in terms of their clinical performance. We believe that this information will be useful for those who have already adopted these platforms or are seeking to implement high-throughput RT-PCR testing to stem the SARS-CoV-2 pandemic.

Bacteremia and Blood Culture Utilization during COVID-19 Surge in New York City.

A surge of patients with coronavirus disease 2019 (COVID-19) presenting to New York City hospitals in March 2020 led to a sharp increase in blood culture utilization, which overwhelmed the capacity of automated blood culture instruments. We sought to evaluate the utilization and diagnostic yield of blood cultures during the COVID-19 pandemic to determine prevalence and common etiologies of bacteremia and to inform a diagnostic approach to relieve blood culture overutilization. We performed a retrospective cohort analysis of 88,201 blood cultures from 28,011 patients at a multicenter network of hospitals within New York City to evaluate order volume, positivity rate, time to positivity, and etiologies of positive cultures in COVID-19. Ordering volume increased by 34.8% in the second half of March 2020 compared to the level in the first half of the month. The rate of bacteremia was significantly lower among COVID-19 patients (3.8%) than among COVID-19-negative patients (8.0%) and those not tested (7.1%) (P < 0.001). COVID-19 patients had a high proportion of organisms reflective of commensal skin microbiota, which, when excluded, reduced the bacteremia rate to 1.6%. More than 98% of all positive cultures were detected within 4 days of incubation. Bloodstream infections are very rare for COVID-19 patients, which supports the judicious use of blood cultures in the absence of compelling evidence for bacterial coinfection. Clear communication with ordering providers is necessary to prevent overutilization of blood cultures during patient surges, and laboratories should consider shortening the incubation period from 5 days to 4 days, if necessary, to free additional capacity.

SARS-CoV-2 Viral Load Predicts Mortality in Patients with and without Cancer Who Are Hospitalized with COVID-19.

Patients with cancer may be at increased risk of severe coronavirus disease 2019 (COVID-19), but the role of viral load on this risk is unknown. We measured SARS-CoV-2 viral load using cycle threshold (CT) values from reverse-transcription polymerase chain reaction assays applied to nasopharyngeal swab specimens in 100 patients with cancer and 2,914 without cancer who were admitted to three New York City hospitals. Overall, the in-hospital mortality rate was 38.8% among patients with a high viral load, 24.1% among patients with a medium viral load, and 15.3% among patients with a low viral load (p < 0.001). Similar findings were observed in patients with cancer (high, 45.2% mortality; medium, 28.0%; low, 12.1%; p = 0.008). Patients with hematologic malignancies had higher median viral loads (CT = 25.0) than patients without cancer (CT = 29.2; p = 0.0039). SARS-CoV-2 viral load results may offer vital prognostic information for patients with and without cancer who are hospitalized with COVID-19.

Safety, tolerability, and clinical outcomes of hydroxychloroquine for hospitalized patients with coronavirus 2019 disease.

BACKGROUND: Severe acute respiratory coronavirus 2 (SARS-CoV-2) has caused a devastating worldwide pandemic. Hydroxychloroquine (HCQ) has in vitro activity against SARS-CoV-2, but clinical data supporting HCQ for coronavirus disease 2019 (COVID-19) are limited. METHODS: This was a retrospective cohort study of hospitalized patients with COVID-19 who received ≥1 dose of HCQ at two New York City hospitals. We measured incident Grade 3 or 4 blood count and liver test abnormalities, ventricular arrhythmias, and vomiting and diarrhea within 10 days after HCQ initiation, and the proportion of patients who completed HCQ therapy. We also describe changes in Sequential Organ Failure Assessment hypoxia scores between baseline and day 10 after HCQ initiation and in-hospital mortality. RESULTS: None of the 153 hospitalized patients with COVID-19 who received HCQ developed a sustained ventricular tachyarrhythmia. Incident blood count and liver test abnormalities occurred in <15% of patients and incident vomiting or diarrhea was rare. Eighty-nine percent of patients completed their HCQ course and three patients discontinued therapy because of QT prolongation. Fifty-two percent of patients had improved hypoxia scores 10 days after starting HCQ. Thirty-one percent of patients who were receiving mechanical ventilation at the time of HCQ initiation died during their hospitalization, compared to 18% of patients who were receiving supplemental oxygen but not requiring mechanical ventilation, and 8% of patients who were not requiring supplemental oxygen. Co-administration of azithromycin was not associated with improved outcomes. CONCLUSIONS: HCQ appears to be reasonably safe and tolerable in most hospitalized patients with COVID-19. However, nearly one-half of patients did not improve with this treatment, highlighting the need to evaluate HCQ and alternate therapies in randomized trials.