Observational study of changes in utilization and outcomes in mechanical ventilation in COVID-19.

BACKGROUND: The role of non-invasive ventilation (NIV) in severe COVID-19 remains a matter of debate. Therefore, the utilization and outcome of NIV in COVID-19 in an unbiased cohort was determined. AIM: The aim was to provide a detailed account of hospitalized COVID-19 patients requiring non-invasive ventilation during their hospital stay. Furthermore, differences of patients treated with NIV between the first and second wave are explored.
METHODS: Confirmed COVID-19 cases of claims data of the Local Health Care Funds with non-invasive and/or invasive mechanical ventilation (MV) in the spring and autumn pandemic period in 2020 were comparable analysed.
RESULTS: Nationwide cohort of 17.023 cases (median/IQR age 71/61-80 years, 64% male) 7235 (42.5%) patients primarily received IMV without NIV, 4469 (26.3%) patients received NIV without subsequent intubation, and 3472 (20.4%) patients had NIV failure (NIV-F), defined by subsequent endotracheal intubation. The proportion of patients who received invasive MV decreased from 75% to 37% during the second period. Accordingly, the proportion of patients with NIV exclusively increased from 9% to 30%, and those failing NIV increased from 9% to 23%. Median length of hospital stay decreased from 26 to 21 days, and duration of MV decreased from 11.9 to 7.3 days. The NIV failure rate decreased from 49% to 43%. Overall mortality increased from 51% versus 54%. Mortality was 44% with NIV-only, 54% with IMV and 66% with NIV-F with mortality rates steadily increasing from 62% in early NIV-F (day 1) to 72% in late NIV-F (>4 days).
CONCLUSIONS: Utilization of NIV rapidly increased during the autumn period, which was associated with a reduced duration of MV, but not with overall mortality. High NIV-F rates are associated with increased mortality, particularly in late NIV-F.

Introduction

Within one year, the SARS-CoV-2 pandemic has affected more than 235 million people worldwide (https://coronavirus.jhu.edu/map.html). Mortality rates of patients requiring ICU treatment are ranging up to over 50% [1-4], depending on the severity of respiratory failure and response to treatment, but also influenced by age, comorbidities and a ceiling of therapeutic interventions [1, 2, 5–7].

Mechanical ventilation (MV) is a life-saving option in severe COVID-19 cases, but mortality rates in patients on MV remain high [4, 5, 8]. Non-invasive ventilation (NIV) is suggested to reduce the complications of invasive MV. The use of noninvasive respiratory support in acute respiratory failure due to viral infection is still debated [9, 10]. For COVID-19 patients, current guidelines recommend stepping up to NIV when oxygenation worsens during oxygen therapy, and to consider intubation if PaO2/FiO2 is decreased below 150 mmHg [11-13] or the clinical presentation of the patients has worsened [11, 13–17], despite international guidelines being still inhomogeneous in recommendations [18]. In Germany, detailed epidemiological data about what types of interfaces are used are not available, but face masks are the most commonly used interface in acute respiratory failure in clinical practice in Europe [19], while helmets are only rarely used by very few experienced centers. However, global current practices of MV widely differ, also depending on COVID-19-associated limited resources [4, 20, 21]. Therefore, the role of NIV remains a matter of uncertainty and discussion, especially with regard to the balance between the NIV benefits and the risk of NIV failure (NIV-F). The mortality of patients receiving NIV was in a wide range up to 45% [1, 22]. In contrast to this, the mortality rate in patients with NIV-F ranged between 35% and 74% [22-24]. Hence, interpretation of data and obtaining conclusive strategies concerning the optimal and individual timing of intubation remain uncertain [25].

Therefore, the aim of the current study was to determine detailed characteristics and outcomes of 7,490 hospitalized COVID-19 patients with MV on the ICU in a large, unselected and unbiased cohort of patients with confirmed COVID-19 in one of the least resource limited health care systems [26], particularly focusing on patients requiring invasive MV or NIV with specific emphasis on NIV-F. Furthermore, we explored the changes between the first spring and second autumn/winter period.

Data and methods

The inpatient data of the general local health insurance funds, which cover around a third of the German population, were analyzed. In general, this is a retrospective analysis of claims data from this registry. Data extraction was done by the scientific institute of the health care insurance, whereas analysis was done by the author group. It is an administrative data set containing patient information like age, gender, diagnosis and procedure codes. However, detailed medical information such as laboratory data is not recorded. All cases were included for which admission and discharge dates as well as diagnoses and procedures were coded. Only patients with laboratory-confirmed SARS-CoV-2 infection (diagnosis code U07.1!) were included. The patients were at least 18 years old and were admitted to hospital between February 1, 2020 and November 30, 2020.

The original data structure is at the case level, i.e. insured persons who were transferred to another hospital during their hospital stay appear several times in the data set. Therefore, cases who were transferred during their hospital stay (discharge date of one hospital corresponds to the admission date of another hospital) were merged. Thus, the current analysis was performed at the patient level. The following OPS codes from the German DRG systematic coding were analyzed: 8–701, 8–704, 5–311, 5–313 and 8–706. For continuous variables, we report means with SDs and medians with IQRs. For categorical variables, we report absolute numbers and percentages.

The primary analysis includes all patients with mechanical ventilation, either non-invasive or invasive on the ICU without any missing, but secondary analysis focuses on patients with MV for more than 6 hours, i.e. invasive MV or NIV. These patients were divided into three subgroups: 1) patients with primary invasive MV without any NIV attempt preceding intubation, 2) patients with NIV exclusively, who have not been escalated to intubation, and 3) those with NIV-F, defined by endotracheal intubation following NIV. In the last group, a procedure code for both NIV and invasive MV was assigned. If invasive MV was started on the day following NIV, the patient was assigned to the NIV-F group. If invasive ventilation was started on the same day as NIV initiation, the patient was assigned to the invasive MV. This definition might help to distinguish real NIV establishment from NIV as short preoxygenation before intubation. Both, patients with less than 6 documented hours of ventilation (n = 695; Table 1) and patients with more than 6 documented hours of ventilation but without a corresponding procedure code for NIV or invasive ventilation (n = 305; Table 1) were not assigned to the three subgroups. With the inclusion of the procedural data, it was possible to roughly determine the NIV duration when switching from NIV to invasive MV.

Table 1

Patient characteristics comparing the spring and autumn period, ECMO = extracorporeal membrane oxygenation.

The Elixhauser Comorbidity Index is a method of categorizing comorbidities of patients based on the (ICD) diagnosis codes in administrative data.

	Patients by month of hospital admission
Variable	Total	Admission between February 2020 and May 2020	Admission between October 2020 and February 2021
Number of patients	17023	2376	13998
Age (years)
Mean (SD)	69.3 (13.1)	68.1 (13.3)	69.7 (12.9)
Median (IQR)	71.0 (61.0, 80.0)	70.0 (60.0, 79.0)	71.0 (62.0, 80.0)
18–49 years	1,304 (7.7%)	204 (8.6%)	1,010 (7.2%)
50–59 years	2,328 (13.7%)	380 (16.0%)	1,840 (13.1%)
60–69 years	4,152 (24.4%)	556 (23.4%)	3,418 (24.4%)
70–79 year	4,847 (28.5%)	726 (30.6%)	3,956 (28.3%)
≥ 80 years	4,392 (25.8%)	510 (21.5%)	3,774 (27.0%)
Male	10,926 (64.2%)	1,565 (65.9%)	8,913 (63.7%)
Female	6,097 (35.8%)	811 (34.1%)	5,085 (36.3%)
Elixhauser comorbidities
Hypertension	11,708 (68.8%)	1,545 (65.0%)	9,717 (69.4%)
Diabetes	7,313 (43.0%)	947 (39.9%)	6,097 (43.6%)
Cardiac arrhythmias	7,311 (42.9%)	1,079 (45.4%)	5,966 (42.6%)
Renal failure	4,979 (29.2%)	627 (26.4%)	4,203 (30.0%)
Congestive heart failure	5,871 (34.5%)	786 (33.1%)	4,853 (34.7%)
Chronic pulmonary disease	3,378 (19.8%)	469 (19.7%)	2,780 (19.9%)
Obesity	2,659 (15.6%)	354 (14.9%)	2,201 (15.7%)
Charlson comorbidity index: 0	3,118 (18.3%)	455 (19.1%)	2,537 (18.1%)
Charlson comorbidity index: 1	3,495 (20.5%)	507 (21.3%)	2,858 (20.4%)
Charlson comorbidity index: 2	2,814 (16.5%)	425 (17.9%)	2,285 (16.3%)
Charlson comorbidity index: 3–4	4,256 (25.0%)	558 (23.5%)	3,564 (25.5%)
Charlson comorbidity index: ≥ 5	3,340 (19.6%)	431 (18.1%)	2,754 (19.7%)
Patients transferred between hospitals	5,543 (32.6%)	856 (36.0%)	4,373 (31.2%)
Length of hospital stay (days)
Mean (SD)	32.9 (33.8)	37.2 (38.7)	31.1 (31.0)
Median (IQR)	22.0 (13.0, 41.0)	26.0 (13.0, 49.0)	21.0 (12.0, 38.0)
Ventilation (days)
Mean (SD)	13.9 (17.7)	17.4 (19.2)	13.1 (17.2)
Median (IQR)	8.0 (2.3, 18.2)	11.9 (4.8, 23.4)	7.3 (2.1, 17.2)
Tracheostomy	4,017 (23.6%)	747 (31.4%)	3,061 (21.9%)
ECMO	1,129 (6.6%)	190 (8.0%)	864 (6.2%)
Dialysis	3,781 (22.2%)	723 (30.4%)	2,891 (20.7%)
Type of ventilation
Invasive ventilation only (IMV)	7,235 (42.5%)	1,772 (74.6%)	5,105 (36.5%)
Non-invasive ventilation only (NIV)	4,469 (26.3%)	221 (9.3%)	4,125 (29.5%)
Non-invasive ventilation failure (NIV-F)	3,472 (20.4%)	214 (9.0%)	3,156 (22.5%)
Duration of ventilation between 1–6 hours	1,152 (6.8%)	85 (3.6%)	1,026 (7.3%)
No ventilation procedure code	695 (4.1%)	84 (3.5%)	586 (4.2%)
In-hospital mortality	9,066 (53.3%)	1,204 (50.7%)	7,607 (54.3%)

The study was approved by the local ethical committee (University Witten/Herdecke, 92/2020).

Findings

Between February 1 2020 and February 28 2021, 16.328 hospitalized Covid-19 patients received MV (Table 1). Data of patients treated during the summer months (June to September) is not shown separately in the table due to the relatively low number. Age distribution, sex and the frequency of comorbidities show only slight differences when comparing the two periods of the pandemic as shown in Table 1. The overall median length of hospital stay has decreased from 26 days during the first wave of the pandemic to 21 days during the second wave. This also applies to the overall duration of MV, which decreased from 11.9 to 7.3 days, respectively.

A major difference between the spring and autumn period of the pandemic refers to the application of the different MV modalities. During the first period, 755% of the patients received invasive MV directly without having previously received NIV as a first escalation step. In contrast, only 37% received immediate invasive MV during the second pandemic wave. Consequently, more patients were escalated from oxygen therapy to NIV during the second period (Table 1) with both patients successfully treated with NIV increasing from 9% to 30% and those with NIV-F increasing from 9% to 23%. However, the overall NIV-F rate decreased from 49% (214 of 435) to 43% (3156 of 7281).

The overall mortality rate of patients receiving any form of MV in the first and second wave of the pandemic increased from 51% and 54%. Overall mortality rates were lower for patients receiving NIV only (44%) compared to those with invasive MV only (54%), as illustrated in more detail in Fig 1A. However, mortality rates of patients with NIV-F were highest (66%). Of note, the mortality rate in patients with NIV-F increased steadily, from 32% in patients with NIV-F on the first day to 72% in those with NIV-F on day 5 or later (Fig 1B).

Fig 1

A. In-hospital mortality by type of ventilation. IMV = invasive mechanical ventilation (n = 3851), NIV = non-invasive-ventilation failure (n = 1614) and NIV-F = non-invasive-ventilation failure (n = 1247). B. In-hospital mortality of non-invasive-ventilation failure (NIV-F, n = 1247) by day of intubation.

Overall, 7.941 patients had initially received NIV, with 3.472 having failed (NIV-F rate 44%) (Table 2). The highest proportion of NIV-F was found in the age group between 60–79 years. The NIV-F rate was lower in women (31%) than in men (69%). There was no clear trend for the influence of comorbidities both on the decision to intubate the patient immediately and on the risk of NIV-F, i.e. the proportion of patients with a specific comorbidity was similar in both groups. One exception refers to cardiac arrhythmias, which were lowest in patients successfully treated with NIV (Table 2).

Table 2

Patient characteristics by type of ventilation.

	Invasive ventilation only (IMV)	Non-invasive ventilation only (NIV)	Non-invasive ventilation failure (NIV-F)	Non-invasive ventilation failure by day of failure
				First day	Second day	Third or fourth day	Fifth day or later
Number of patients	7235	4469	3472	1213	591	684	984
Age (years)
Mean (SD)	68.0 (12.9)	71.2 (13.5)	68.6 (11.9)	67.9 (12.5)	68.6 (12.3)	68.9 (11.4)	69.1 (11.0)
Median (IQR)	70.0 (60.0, 78.0)	74.0 (62.0, 82.0)	70.0 (62.0, 78.0)	69.0 (60.0, 78.0)	70.0 (61.0, 78.0)	71.0 (62.0, 78.0)	70.0 (62.0, 78.0)
18–49 years	613 (8.5%)	325 (7.3%)	228 (6.6%)	101 (8.3%)	42 (7.1%)	38 (5.6%)	47 (4.8%)
50–59 years	1,095 (15.1%)	541 (12.1%)	470 (13.5%)	176 (14.5%)	78 (13.2%)	88 (12.9%)	128 (13.0%)
60–69 years	1,904 (26.3%)	886 (19.8%)	996 (28.7%)	336 (27.7%)	173 (29.3%)	197 (28.8%)	290 (29.5%)
70–79 year	2,135 (29.5%)	1,131 (25.3%)	1,102 (31.7%)	358 (29.5%)	178 (30.1%)	230 (33.6%)	336 (34.1%)
≥ 80 years	1,488 (20.6%)	1,586 (35.5%)	676 (19.5%)	242 (20.0%)	120 (20.3%)	131 (19.2%)	183 (18.6%)
Male	4,687 (64.8%)	2,725 (61.0%)	2,405 (69.3%)	816 (67.3%)	418 (70.7%)	475 (69.4%)	696 (70.7%)
Female	2,548 (35.2%)	1,744 (39.0%)	1,067 (30.7%)	397 (32.7%)	173 (29.3%)	209 (30.6%)	288 (29.3%)
Elixhauser comorbidities
Hypertension	4,913 (67.9%)	3,129 (70.0%)	2,482 (71.5%)	839 (69.2%)	428 (72.4%)	476 (69.6%)	739 (75.1%)
Diabetes	3,182 (44.0%)	1,889 (42.3%)	1,556 (44.8%)	543 (44.8%)	258 (43.7%)	302 (44.2%)	453 (46.0%)
Cardiac arrhythmias	3,304 (45.7%)	1,619 (36.2%)	1,641 (47.3%)	557 (45.9%)	290 (49.1%)	327 (47.8%)	467 (47.5%)
Renal failure	2,006 (27.7%)	1,430 (32.0%)	931 (26.8%)	323 (26.6%)	172 (29.1%)	185 (27.0%)	251 (25.5%)
Congestive heart failure	2,494 (34.5%)	1,510 (33.8%)	1,195 (34.4%)	398 (32.8%)	212 (35.9%)	238 (34.8%)	347 (35.3%)
Chronic pulmonary disease	1,351 (18.7%)	950 (21.3%)	716 (20.6%)	230 (19.0%)	118 (20.0%)	151 (22.1%)	217 (22.1%)
Obesity	1,220 (16.9%)	610 (13.6%)	624 (18.0%)	232 (19.1%)	111 (18.8%)	126 (18.4%)	155 (15.8%)
Charlson comorbidity index: 0	1,192 (16.5%)	953 (21.3%)	613 (17.7%)	226 (18.6%)	109 (18.4%)	115 (16.8%)	163 (16.6%)
Charlson comorbidity index: 1	1,479 (20.4%)	959 (21.5%)	722 (20.8%)	252 (20.8%)	104 (17.6%)	151 (22.1%)	215 (21.8%)
Charlson comorbidity index: 2	1,179 (16.3%)	725 (16.2%)	620 (17.9%)	220 (18.1%)	107 (18.1%)	124 (18.1%)	169 (17.2%)
Charlson comorbidity index: 3–4	1,854 (25.6%)	1,050 (23.5%)	890 (25.6%)	319 (26.3%)	157 (26.6%)	185 (27.0%)	229 (23.3%)
Charlson comorbidity index: ≥ 5	1,531 (21.2%)	782 (17.5%)	627 (18.1%)	196 (16.2%)	114 (19.3%)	109 (15.9%)	208 (21.1%)
Patients transferred between hospitals	2,963 (41.0%)	800 (17.9%)	1,397 (40.2%)	501 (41.3%)	228 (38.6%)	267 (39.0%)	401 (40.8%)
Length of hospital stay (days)
Mean (SD)	41.5 (40.1)	20.4 (16.8)	37.6 (33.6)	35.6 (32.3)	35.6 (32.0)	34.7 (29.9)	43.2 (37.7)
Median (IQR)	29.0 (16.0, 53.5)	16.0 (10.0, 26.0)	27.0 (16.0, 47.0)	26.0 (15.0, 45.0)	26.0 (16.0, 44.5)	25.0 (15.0, 44.2)	30.0 (19.0, 54.0)
Ventilation (days)
Mean (SD)	19.0 (19.2)	4.1 (4.7)	21.6 (20.2)	20.1 (19.5)	20.6 (19.5)	21.5 (20.3)	24.2 (21.3)
Median (IQR)	13.3 (6.3, 25.1)	2.7 (1.2, 5.3)	15.8 (9.2, 27.5)	14.3 (8.2, 25.7)	15.5 (8.9, 25.3)	15.6 (9.3, 26.4)	18.0 (10.5, 31.0)
Tracheostomy	2,705 (37.4%)	0 (0.0%)	1,312 (37.8%)	437 (36.0%)	216 (36.5%)	256 (37.4%)	403 (41.0%)
ECMO	616 (8.5%)	14 (0.3%)	479 (13.8%)	155 (12.8%)	71 (12.0%)	90 (13.2%)	163 (16.6%)
Dialysis	2,130 (29.4%)	257 (5.8%)	1,188 (34.2%)	396 (32.6%)	201 (34.0%)	240 (35.1%)	351 (35.7%)
In-hospital mortality	3,874 (53.5%)	1,949 (43.6%)	2,306 (66.4%)	746 (61.5%)	382 (64.6%)	470 (68.7%)	708 (72.0%)

ECMO = extracorporeal membrane oxygenation, NIV = non-invasive ventilation, NIV-F = non-invasive ventilation failure.

The duration of MV was clearly dependent on its modality (Table 2). The median duration of MV was 2.7 days in those receiving NIV only but reached 14 days in those who were intubated directly. Of note, patients who were switched from initial NIV to invasive MV following NIV failure spent the longest periods on MV (median 16 days). This trend was also true for the application of ECMO, which was reported in 14% of NIV-F patients, compared to 9% in patients who were intubated without having initially received NIV. Importantly, the proportion of patients with late NIV-F (after 5 days or more of NIV followed by intubation) substantially increased during the second wave, as displayed in Fig 2.

Fig 2

Time distribution of NIV failure (NIV-F) by day, comparing spring and autumn period.

Discussion

The current analysis of 17.023 patients represents the largest case series of COVID-19 patients requiring NIV or invasive MV and shows significant differences between the spring and autumn/winter periods 2020/21 with regard to the modality of MV. The main findings are as follows: Firstly, there was a significant increase in the utilization of NIV in Germany during the second period. Accordingly, the proportion of patients with acute respiratory failure who were directly intubated decreased from 75% to 37%. This was associated with a reduced overall duration of MV, and length of hospital stay. Secondly, the NIV-F rate was still high, even though there was a trend for a lower NIV-F rate during the second period (42%) compared to the first period (49%). Thirdly, the overall mortality rate in patients requiring MV remains high at 54%. Fourthly, NIV-F was associated with an increased ECMO utilization, increased overall duration of MV and increased mortality, and this was particularly true for late NIV-F occurring 5 days or later following NIV initiation.

Several clinical considerations can be derived from the current findings. Most importantly, the present analysis shows that NIV has been clearly established in the treatment of severe respiratory failure attributable to COVID-19 in a real-life setting without resource limitations, since this is shown on a daily basis in the nationwide ICU registry, which counts all free and occupied ICU beds and COVID patients respectively (www.intensivregister.de). Thereby, in 2020 the overall duration of MV and hospital stay could be shortened. The decreasing NIV-F rate also suggests a learning curve that has occurred over the course of the last year, but may also be related to treatment successes outside MV, such as corticosteroids or prone positioning even with NIV at least in some centers [27, 28]. Remarkably, also the need of renal replacement therapy markedly dropped. Here, also steroids and/ or optimized treatment strategies, especially on mechanical ventilation may have attributed to this finding.

The present analysis, however, also demonstrates that clinicians should apply NIV cautiously as NIV-F continues to occur frequently, which is associated with increased mortality. Although reasons for intubation after NIV were not examined in this study, we did observe that a substantial proportion of patients initially treated with NIV progressed to receipt of IMV. The rather short median duration of NIV of 2.7 days in those patients successfully treated by NIV suggests that early improvements in respiratory function following NIV identify those patients who have been successfully treated and do not need intubation. In contrast, a longer duration of NIV, particularly exceeding 3–5 days, increases the likelihood of NIV-F, which is associated with an increased mortality. However, although administrative data in general contain no information e.g. about do not intubated or do not resuscitate order, the definition of NIV failure remained the same over the entire period of the study.

There are many other reports in the literature also showing the potential of NIV in the treatment of COVID-19-associated respiratory failure [29-31], and this might also have encouraged clinicians to more frequently and extensively apply NIV in this setting. In these reports, NIV was reasonably used outside the ICU, in part aimed at overcoming the shortage of ICU capacities [31]. Another rationale to use NIV as long as possible, also in the ICU setting, is aimed at avoiding intubation and intubation-related complications, most importantly lung injury related to invasive MV and infectious complications [9].

In this context, however, clinicians are likely to be less aware of a phenomenon related to maintained spontaneous breathing, which is labeled as patient self-inflicted lung injury (P-SILI) [32]. In short, initial lung injury related to COVID-19 is perpetually maintained and even aggravated as a consequence of a vicious circle that includes the sequence of capillary leakage, pulmonary edema, impaired gas exchange and respiratory mechanics, subsequent increase in respiratory drive followed by increased pleural pressure swings, which eventually lead to capillary leakage again if lowering of pleural pressure exceeds the intravascular pressure decrease [33, 34]. Even though the current data does not provide evidence for P-SILI in those patients having failed NIV, this phenomenon might, nevertheless, explain why outcome is severely reduced in patients spending longer durations on NIV, which eventually fails. Of note, the current data demonstrate, that women show independently of the first or second wave a less severe course of the disease.

Limitations

There are some important limitations of the present data, which need to be addressed in the context of data interpretation. First, all data refer to the coding of diseases (ICD) and procedures (OPS) in the context of remuneration. Thus, patients were not studied directly. Therefore, several important data are missing, i.e. disease severity related to the PaO2/FiO2 ratio, intubation criteria, ventilator settings/equipment and oxygen flow rates including information on the response to treatment, “do-not-intubate” orders or details on medication.

Second, the analysis includes only data from one health care insurance company. However, this is the largest insurance and accounts for about 1/3 of the total population, providing a large representative sample for the German population. Third, NIV as defined for the reimbursement system in Germany excludes high-flow oxygen treatment (HFOT) and continuous positive airway pressure (CPAP) and can only be coded if the level of pressure support exceeds 5 cm H2O. Therefore, this analysis focused on non-invasive and invasive ventilation, which are both very accurately documented since reimbursement in ICU medicine in Germany mainly depends on MV. In addition, the German guidelines have recommended using HFOT as first escalation step when oxygen treatment is insufficient, while CPAP and NIV form the following escalation steps. Thus, NIV in the present analysis represents a rather selected group of patients, and this group may not be compared to studies from other countries without considering this aspect. Furthermore, patients with less than 6 documented hours of ventilation and patients with more than 6 documented hours of ventilation but without a corresponding procedure code for NIV or invasive ventilation were not assigned to the three subgroups. However, the patient number is low in these groups. Finally, since many factors are likely undetectable in the administrative dataset, unmeasured confounding and confounding by indication and over-or under-reporting or misclassification of cases remain additional major limitations of the study. These challenges are best addressed with a multicenter and multinational clinical trial that randomizes patients to NIV vs. IMV, with clear clinical criteria to standardize crossover to IMV.

Conclusions

The utilization of NIV rapidly increased during the autumn/winter period compared to the spring period 2020 of the COVID-19 pandemic in Germany. This was associated with an overall reduced duration of MV, and length of hospital stay. However, the current data do not explain in detail the reasons of mortality, since also other treatment modifications may have contributed to the outcome. Despite this, overall mortality of patients receiving MV due to COVID-19-associated respiratory failure remained high at 53%. Patients successfully treated with NIV had lower mortality rates than those who were intubated directly, but those failing NIV had a higher mortality rate, respectively, and this became even more predominant in late NIV failure. Thus, the current study shows the increasing role of NIV in the context of ICU medicine related to COVID-19 during the second wave and, may also emphasize on its risks. Prompt identification of patients failing an NIV approach is mandatory to avoid harmful delays and very poor outcomes. Given these findings, there is a need for prospective randomized controlled trials that focus on the most reasonable indications for initiation of NIV as well as timely subsequent intubation in case of NIV failure in COVID-19 patients.

28 Aug 2021

PONE-D-21-23922

Observational study of changes in utilization and outcomes in non-invasive ventilation in COVID-19

PLOS ONE

Dear Dr. Karagiannidis,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

ACADEMIC EDITOR: I support the comments from the Reviewers, that are two experts in the field. Specifically, more details on the methods and on statistical plan and approach should be provided to re-evaluate the paper.

Please submit your revised manuscript by Oct 12 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Andrea Cortegiani, M.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following in the Competing Interests section:

"Dr. Karagiannidis reports personal fees from Maquet, personal fees from Xenios, personal fees from Bayer, non-financial support from Speaker of the German register of ICUs, grants from German Ministry of Research and Education, during the conduct of the study. Dr. Hentschker has nothing to disclose. Dr. Westhoff has nothing to disclose. Dr. Weber-Carstens has nothing to disclose. Dr. Janssens has nothing to disclose. Dr. Kluge reports non-financial support from Ambu, ETView Ltd, Fisher & Paykel and Xenios., grants from Daiichi Sankyo, Pfizer, personal fees from Astra, C.R. Bard, Baxter, Biotest, Cytosorbents, Fresenius, Gilead, MSD, Pfizer, Philips, ZOLL, personal fees and other from Bayer, Fresenius, Gilead, MSD und Pfizer, outside the submitted work. MP reports no conflicts of interests in regard to the manuscript, lecture fees from Boehringer, Novartis, Astra_Zeneca, Roche and fees for advisory board meetings from Boehringer, Novartis, Roche Current president of the German Society of Pneumology. Dr. Spies reports grants from Public Grants, grants from IIT grants from companies, other from Meeting support from companies (e.g. for the Leopoldina 2020 meeting), outside the submitted work; In addition, Dr. Spies has a patent EEG monitoring licensed, and a patent Ceilings licensed. Dr. Welte reports grants from German Minstry of Research and Education, during the conduct of the study."

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests).  If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

3. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I would like to thank the Editor for the chance to review this work by Karagiannidis and co-workers.

The authors present a considerable amount of interesting data regarding the changes in use of mechanical ventilation in COVID-19 patients in two pandemic waves during 2020. They report the use of NIV rapidly increased during the second wave without change in overall mortality.

General comment

The question is appealing as the use of NIV in hypoxic patients (including COVID-19) still remains a hot topic. My personal opinion is that, although data are interesting given the considerable number of patients enrolled, several methodological issues negatively affect the quality of the work. In particular the lack of a pre-specified statistical approach to this huge amount of data and the absence of a solid method to account for confounders in the analysis, do not allow to draw conclusions.

Specific comments

Title

The title refers to the change in use and outcome of NIV. However, a consistent amount of data and analyses focus on the use of invasive mechanical ventilation in patients that did not receive NIV trial. This should be reported in the title. (i.e. Observational study of changes in utilization and outcomes in mechanical ventilation in COVID-19) otherwise analysis should be limited only to NIV use over time.

Introduction

Line 129. Is the analysis only limited to the intensive care setting? During the pandemic NIV has been extensively used even outside the ICU. It would be important to know if these data refer to the use of NIV in general or only limited to the ICU setting (where technological monitoring, trained nursing and specific critical care skills are far more granted).

Methods

The methods section should be expanded to illustrate the pre-specified analyses performed once data had been categorized (invasive MV, NIV, NIV-F). What statistical approach was applied? Which inferential methodology? This information is essential to understand the nature of this investigation and the reliability of findings.

Line 142. I am not sure this would be a reliable method to assess NIV duration as it may be discontinued over time and then resumed for clinical worsening. Consider adding this point in the limitations section.

Discussion

As reported in the limitations section, one of the major limitations is the lack of a pre-specified definition of NIV failure. Was this definition the same in the two pandemic periods?More details should be provided on this topic.

How did the authors account for NIV patients with ceiling of escalation to intensive care? This subgroup of patients might have affected the results of NIV mortality.

Conclusion

I am not sure conclusions regarding mortality are sufficiently supported by the presented results as many confounding factors were not included in the analysis.

Reviewer #2: GENERAL COMMENTS

Thank you for allowing me to review this interesting manuscript. This is a retrospective study of administrative (healthcare insurance) database analysis showing NIV utilization in patients with SARS-CoV-2 infection in Germany. The authors concluded that successful NIV lowered mortality rates, but NIV failure was associated with greater mortality risk. Therefore, prompt identification of those failing an NIV approach is mandatory to avoid harmful delays and very poor outcomes. The manuscript deals with a clinically relevant topic, given that the role of NIV is still debated in COVID-19 ARDS.

SPECIFIC COMMENTS

The quality of written English is acceptable.

Major comment

Abstract:

1) Aim: Please specify that you also want to explore differences of patients treated with NIV between first and second wave

Introduction:

The background is concise and informative. However, I have some suggestions:

1) Line 89 this statement needs a reference.

2) Line 94 – 95 “Non-invasive ventilation (NIV) is suggested to reduce the complications of invasive MV(9)”. I suggest to state that the use of noninvasive respiratory support in acute respiratory failure due to viral infection is still debated quoting as reference a systematic review to help the readers better understand the context (e.g. doi: 10.23736/S0375-9393.20.14785-0)

3) Line 95-98: I suggest to add that Guidelines from different regions on the use of NIV in COVID-19 have been inconsistent and heterogeneous quoting this reference doi: 10.1016/j.ijid.2021.03.078

4) Lines 98-101 you state that “personal communication suggests that face masks are by far the most widely used interfaces”. What do you mean by "personal communication"? Please clarify. This statement needs a reference, as reported, it looks like an opinion. You might state that face masks are the most commonly used interface in acute respiratory failure in clinical practice in Europe, quoting this survey doi:10.1183/09031936.00123509

5) Lines 109-113 Aim: I would add that you also want to explore practice changes between spring and autumn period.

.

Methods:

The methods used are appropriate for the retrospective design of the study based on data registry. Study question is clearly stated and clinically relevant. However, I have some remarks:

1) Please specify that this is a retrospective analysis of claim data from registry.

2) Who did perform data extraction from the insurance database? research staff of the insurance company or study investigator? Please specify.

3) Please clarify in the methods section how you did account for comorbidities. I see in Table 1 that you used "Elixhauser comorbidities", please add this information in the methods section and provide a reference for "Elixhauser comorbidities". Moreover, I would suggest also adding the Charlson comorbidity index, which is much more used in clinical practice and in clinical research and it may be more helpful for the reader to interpret the data from the study.

4) Please specify the ICD procedure code used to identify the use of NIV and IMV.

5) In your search, was COVID-19 the primary admission diagnosis code for hospitalization from ICD or one of the secondary admission diagnoses? Please specify this aspect. Selected patients with a different (non-COVID) primary admitting diagnosis might also require NIV or IMV and develop COVID during hospitalization (especially in the first wave of pandemic).

RESULTS

1) A study flowchart to illustrate included/excluded patients would be helpful to the reader to explain the flow of patients in the study

2) Table 1: I would add the overall NIV-F rate number of patients who received NIV prior to ICU admission (Noninvasive ventilation on ICU admission)

DISCUSSION

The discussion is balanced. References are relevant and updated. Limitations of the study are correctly addressed by the authors and discussed, but I would recommend emphasizing some aspects:

1) Please emphasize that, as with all data routinely collected for other purposes (health insurances), the accuracy and completeness of the information may be compromised due to over-or under-reporting or misclassification of cases.

2) Ventilation days, tracheostomy, and dialysis differed between the 2 periods (nearly 10% less during the second wave). Were patients of the second wave less severe? Please discuss this point.

3) How would you explain the lack of improvement in intensive care mortality or ventilated patients' prognosis despite the decreasing NIV-F rate in the second wave?

4) How would you explain the change in clinical practice on NIV use between the two periods? Please discuss this aspect.

5) Was awake prone positioning during NIV part of the usual clinical practice in Germany? If yes, please add this intervention in lines 213-215 "The decreasing NIV-F rate also suggests a learning curve that has occurred over the course of the last year, but may also be related to treatment successes outside MV, such as corticosteroids (24, 25)".

6) Line 252-253 In the limitation settings you state that is not possible to identify the hospital settings (ICU, intermediate care, COVID-19 wards) but in the introduction lines 109-110 you state that the aim of the study is "to determine detailed characteristics and outcomes of 7,490 hospitalized COVID-19 patients with MV on the ICU", please clarify.

7) I would emphasize more the message that prompt identification of patients failing an NIV approach is mandatory to avoid harmful delays and very poor outcomes.

Minor comments:

1) Table 1 “elixhauser” please report the full definition in the table legend

2) Line 210-211 please check punctuation

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

16 Oct 2021

Response to Reviewer 1: We appreciate the very thoughtful critiques by the reviewers, and we have revised our manuscript accordingly. Furthermore, we updated the data extensively until February 2021, which remarkably increased the patient number of the study from 7000 to 17000.

Reviewer #1: I would like to thank the Editor for the chance to review this work by Karagiannidis and co-workers.

The authors present a considerable amount of interesting data regarding the changes in use of mechanical ventilation in COVID-19 patients in two pandemic waves during 2020. They report the use of NIV rapidly increased during the second wave without change in overall mortality.

General comment

The question is appealing as the use of NIV in hypoxic patients (including COVID-19) still remains a hot topic. My personal opinion is that, although data are interesting given the considerable number of patients enrolled, several methodological issues negatively affect the quality of the work. In particular the lack of a pre-specified statistical approach to this huge amount of data and the absence of a solid method to account for confounders in the analysis, do not allow to draw conclusions.

Thank you very much for your effort. We revised our manuscript accordingly. Furthermore, we updated the patient data and included now all patients from the second wave until February 2021. This increased the patient number substantially.

Specific comments

Title

The title refers to the change in use and outcome of NIV. However, a consistent amount of data and analyses focus on the use of invasive mechanical ventilation in patients that did not receive NIV trial. This should be reported in the title. (i.e. Observational study of changes in utilization and outcomes in mechanical ventilation in COVID-19) otherwise analysis should be limited only to NIV use over time.

Thank you very much for this very important point. We changed the title accordingly.

Introduction

Line 129. Is the analysis only limited to the intensive care setting? During the pandemic NIV has been extensively used even outside the ICU. It would be important to know if these data refer to the use of NIV in general or only limited to the ICU setting (where technological monitoring, trained nursing and specific critical care skills are far more granted).

Our Data are limited to the ICU setting. In Germany only very few patients are treated outside the ICU with NIV. However, some of these patients are treated on so-called “low-care ICU beds”. We added this point to the methods: “The primary analysis includes all patients with mechanical ventilation, either non-invasive or invasive on the ICU without any missing”

Methods

The methods section should be expanded to illustrate the pre-specified analyses performed once data had been categorized (invasive MV, NIV, NIV-F). What statistical approach was applied? Which inferential methodology? This information is essential to understand the nature of this investigation and the reliability of findings.

We added some more specific comments to the Methods part now: “Therefore, cases who were transferred during their hospital stay (discharge date of one hospital corresponds to the admission date of another hospital) were merged. Thus, the current analysis was performed at the patient level. The following OPS codes from the German DRG systematic coding were analyzed: 8-701, 8-704, 5-311, 5-313 and 8-706. For continuous variables, we report means with SDs and medians with IQRs. For categorical variables, we report absolute numbers and percentages.”

Line 142. I am not sure this would be a reliable method to assess NIV duration as it may be discontinued over time and then resumed for clinical worsening. Consider adding this point in the limitations section.

We added this point accordingly in the limitation section: “Furthermore, patients with less than 6 documented hours of ventilation and patients with more than 6 documented hours of ventilation but without a corresponding procedure code for NIV or invasive ventilation were not assigned to the three subgroups. However, the patient number is low in these groups”

Discussion

As reported in the limitations section, one of the major limitations is the lack of a pre-specified definition of NIV failure. Was this definition the same in the two pandemic periods? More details should be provided on this topic.

We clarified this topic in the discussion section: “However, although administrative data in general contain no information e.g. about do not intubated or do not resuscitate order, the definition of NIV failure remained the same over the entire period of the study.”

How did the authors account for NIV patients with ceiling of escalation to intensive care? This subgroup of patients might have affected the results of NIV mortality.

In Germany only few patients with NIV are treated outside the ICU. The reason for it is, that reimbursement is dependent on being on the ICU and then counting the numbers of hours on MV. We would be happy if we may not address this point in the manuscript.

Conclusion

I am not sure conclusions regarding mortality are sufficiently supported by the presented results as many confounding factors were not included in the analysis.

We agree and integrate the following sentence to the conclusion: “However, the current data do not explain in great detail the reasons of mortality, since also other treatment modifications may have contributed to the outcome.”

Response to Reviewer 1: We appreciate the very thoughtful critiques by the reviewers, and we have revised our manuscript accordingly. Furthermore, we updated the data extensively until February 2021, which remarkably increased the patient number of the study from 7000 to 17000.

Reviewer #2: GENERAL COMMENTS

Thank you for allowing me to review this interesting manuscript. This is a retrospective study of administrative (healthcare insurance) database analysis showing NIV utilization in patients with SARS-CoV-2 infection in Germany. The authors concluded that successful NIV lowered mortality rates, but NIV failure was associated with greater mortality risk. Therefore, prompt identification of those failing an NIV approach is mandatory to avoid harmful delays and very poor outcomes. The manuscript deals with a clinically relevant topic, given that the role of NIV is still debated in COVID-19 ARDS.

SPECIFIC COMMENTS

The quality of written English is acceptable.

Major comment

Abstract:

Aim: Please specify that you also want to explore differences of patients treated with NIV between first and second wave

We are thankful for this comment and integrated this comment now: “Furthermore, differences of patients treated with NIV between the first and second wave are explored.”

Introduction:

The background is concise and informative. However, I have some suggestions:

1) Line 89 this statement needs a reference.

We added this reference now: “Within one year, the SARS-CoV-2 pandemic has affected more than 235 million people worldwide (https://coronavirus.jhu.edu/map.html).”

2) Line 94 – 95 “Non-invasive ventilation (NIV) is suggested to reduce the complications of invasive MV(9)”. I suggest to state that the use of noninvasive respiratory support in acute respiratory failure due to viral infection is still debated quoting as reference a systematic review to help the readers better understand the context (e.g. doi: 10.23736/S0375-9393.20.14785-0)

We revised it accordingly: “The use of noninvasive respiratory support in acute respiratory failure due to viral infection is still debated (9){Crimi, 2020 #2307}”

3) Line 95-98: I suggest to add that Guidelines from different regions on the use of NIV in COVID-19 have been inconsistent and heterogeneous quoting this reference doi: 10.1016/j.ijid.2021.03.078

We added this very valuable point now: “despite international guidelines being still inhomogeneous in recommendations (18)”

4) Lines 98-101 you state that “personal communication suggests that face masks are by far the most widely used interfaces”. What do you mean by "personal communication"? Please clarify. This statement needs a reference, as reported, it looks like an opinion. You might state that face masks are the most commonly used interface in acute respiratory failure in clinical practice in Europe, quoting this survey doi:10.1183/09031936.00123509

Thak you very much, we changed this accordingly: “In Germany, detailed epidemiological data about what types of interfaces are used are not available, but face masks are the most commonly used interface in acute respiratory failure in clinical practice in Europe (19), while helmets are only rarely used by very few experienced centers”

5) Lines 109-113 Aim: I would add that you also want to explore practice changes between spring and autumn period.

We changed this accordingly in the aim.

.

Methods:

The methods used are appropriate for the retrospective design of the study based on data registry. Study question is clearly stated and clinically relevant. However, I have some remarks:

1) Please specify that this is a retrospective analysis of claim data from registry.

We added this now to this section: “In general this is a retrospective analysis of claims data from this registry.”

2) Who did perform data extraction from the insurance database? research staff of the insurance company or study investigator? Please specify.

This was clearly specified now: “Data extraction was done by the scientific institute of the health care insurance, whereas analysis was done by the author group”

3) Please clarify in the methods section how you did account for comorbidities. I see in Table 1 that you used "Elixhauser comorbidities", please add this information in the methods section and provide a reference for "Elixhauser comorbidities". Moreover, I would suggest also adding the Charlson comorbidity index, which is much more used in clinical practice and in clinical research and it may be more helpful for the reader to interpret the data from the study.

We now added the Charlson comorbidity index to the table 1 and 2.

4) Please specify the ICD procedure code used to identify the use of NIV and IMV.

This was added to the Method section now:” Therefore, cases who were transferred during their hospital stay (discharge date of one hospital corresponds to the admission date of another hospital) were merged. Thus, the current analysis was performed at the patient level. The following OPS codes from the German DRG systematic coding were analyzed: 8-701, 8-704, 5-311, 5-313 and 8-706. For continuous variables, we report means with SDs and medians with IQRs. For categorical variables, we report absolute numbers and percentages.”

5) In your search, was COVID-19 the primary admission diagnosis code for hospitalization from ICD or one of the secondary admission diagnoses? Please specify this aspect. Selected patients with a different (non-COVID) primary admitting diagnosis might also require NIV or IMV and develop COVID during hospitalization (especially in the first wave of pandemic).

Most patients are admitted because of COVID-19. We did several analyses on this point by grouping the patients with and without primary respiratory insufficiency as far as you can do that from these data. All robustness checks revealed the same outcome. Therefore it seems that this point occurs, but does not really make a difference.2

RESULTS

1) A study flowchart to illustrate included/excluded patients would be helpful to the reader to explain the flow of patients in the study

We generally agree with this remark, but our search strategy was patients with confirmed COVID diagnosis and MV. Therefore from our perspective a flow chart would not really help the reader.

2) Table 1: I would add the overall NIV-F rate number of patients who received NIV prior to ICU admission (Noninvasive ventilation on ICU admission)

Unfortunately we don´t have the data, but we expect only very, very few patients in this group, since NIV is done on high-care or low-care ICUs in Germany, but not outside. This is very different to other countries in the world.

DISCUSSION

The discussion is balanced. References are relevant and updated. Limitations of the study are correctly addressed by the authors and discussed, but I would recommend emphasizing some aspects:

1) Please emphasize that, as with all data routinely collected for other purposes (health insurances), the accuracy and completeness of the information may be compromised due to over-or under-reporting or misclassification of cases.

We added this point to the limitation section: “Finally, since many factors are likely undetectable in the administrative dataset, unmeasured confounding and confounding by indication and over-or under-reporting or misclassification of cases remain additional major limitations of the study.”

2) Ventilation days, tracheostomy, and dialysis differed between the 2 periods (nearly 10% less during the second wave). Were patients of the second wave less severe? Please discuss this point.

This is an important but unknown point. In the international discussion we had on this are two opinions: better treatment and/or steroids. Many countries observed this.We added this now to the discussion: “Remarkably, also the need of renal replacement therapy markedly dropped. Here, also steroids and/ or optimized treatment strategies, especially on mechanical ventilation may have attributed to this finding.”

3) How would you explain the lack of improvement in intensive care mortality or ventilated patients' prognosis despite the decreasing NIV-F rate in the second wave?

This is highly speculative. Overuse of NIV might be one problem, more bacterial superinfection might be another reason. Since we cannot prove it, we would not like to stress this point in the discussion.

4) How would you explain the change in clinical practice on NIV use between the two periods? Please discuss this aspect.

Indeed, the huge increase in NIV usage at the second wave is remarkable. Even though further explanation is needed, there was an increasingly insistent discussion on the rationale of NIV to be used in COVID-19 patients in the ICU setting that has started to emerge with the dwindling of the first pandemic wave.

5) Was awake prone positioning during NIV part of the usual clinical practice in Germany? If yes, please add this intervention in lines 213-215 "The decreasing NIV-F rate also suggests a learning curve that has occurred over the course of the last year, but may also be related to treatment successes outside MV, such as corticosteroids (24, 25)".

We recommended to use prone positioning in NIV in Octobre 2021. However, some centres did that even before. We added this now:” …such as corticosteroids or prone positioning even with NIV at least in some centers”

6) Line 252-253 In the limitation settings you state that is not possible to identify the hospital settings (ICU, intermediate care, COVID-19 wards) but in the introduction lines 109-110 you state that the aim of the study is "to determine detailed characteristics and outcomes of 7,490 hospitalized COVID-19 patients with MV on the ICU", please clarify.

Thanks for this comment. We removed this part from the limitation section, since this is misleading. We can identify the patients by a code for the ICU in Germany. This is what we did in the analysis. Sorry for the confusion.

7) I would emphasize more the message that prompt identification of patients failing an NIV approach is mandatory to avoid harmful delays and very poor outcomes.

We fully agree and added this point to the conclusions:” Prompt identification of patients failing an NIV approach is mandatory to avoid harmful delays and very poor outcomes.”

Minor comments:

1) Table 1 “elixhauser” please report the full definition in the table legend

We integrated this into the table and legend now.

2) Line 210-211 please check punctuation

Thanks for this very careful correction. This was done.

21 Dec 2021

Observational study of changes in utilization and outcomes in mechanical ventilation in COVID-19

PONE-D-21-23922R1

Dear Dr. Karagiannidis,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Andrea Cortegiani, M.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I would like to thank the Authors for their work and the responses give to my concerns.

I do not have further comments to make.

Reviewer #2: The authors addressed all the points raised by the reviewers and should be congratulated for their effort. The manuscript has improved significantly. I have no additional remarks.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Roberto Tonelli

Reviewer #2: No

5 Jan 2022

PONE-D-21-23922R1

Observational study of changes in utilization and outcomes in mechanical ventilation in COVID-19

Dear Dr. Karagiannidis:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Andrea Cortegiani

Academic Editor

PLOS ONE