Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2020 Oct 30;15(10):e0240721. doi: 10.1371/journal.pone.0240721

Factors influencing the bias between blood gas analysis versus central laboratory hemoglobin testing. A secondary analysis of a randomized controlled trial

Linda Tanner 1, Simone Lindau 1, Markus Velten 2, Tobias Schlesinger 3, Maria Wittmann 2, Peter Kranke 3, Kira Berg 1, Florian Piekarski 1, Christoph Füllenbach 1, Suma Choorapoikayil 1, Dirk Hasenclever 4, Kai Zacharowski 1, Patrick Meybohm 1,3,*
Editor: Laura Pasin5
PMCID: PMC7598475  PMID: 33125405

Abstract

Background

Anemia is the most important complication during major surgery and transfusion of red blood cells is the mainstay to compensate for life threating blood loss. Therefore, accurate measurement of hemoglobin (Hb) concentration should be provided in real-time. Blood Gas Analysis (BGA) provides rapid point-of-care assessment using smaller sampling tubes compared to central laboratory (CL) services.

Objective

This study aimed to investigate the accuracy of BGA hemoglobin testing as compared to CL services.

Methods

Data of the ongoing LIBERAL-Trial (Liberal transfusion strategy to prevent mortality and anemia-associated ischemic events in elderly non-cardiac surgical patients, LIBERAL) was used to assess the bias for Hb level measured by BGA devices (ABL800 Flex analyzer®, GEM series® and RapidPoint 500®) and CL as the reference method. For that, we analyzed pairs of Hb level measured by CL and BGA within two hours. Furthermore, the impact of various confounding factors including age, gender, BMI, smoker status, transfusion of RBC, intraoperative hemodilution, and co-medication was elucidated. In order to ensure adequate statistical analysis, only data of participating centers providing more than 200 Hb pairs were used.

Results

In total, three centers including 963 patients with 1,814 pairs of Hb measurements were analyzed. Mean bias was comparable between ABL800 Flex analyzer® and GEM series®: - 0.38 ± 0.15 g/dl whereas RapidPoint 500® showed a smaller bias (-0.09 g/dl) but greater median absolute deviation (± 0.45 g/dl). In order to avoid interference with different standard deviations caused by the different analytic devices, we focused on two centers using the same BGA technique (309 patients and 1,570 Hb pairs). A Bland-Altman analysis and LOWESS curve showed that bias decreased with smaller Hb values in absolute numbers but increased relatively. The smoker status showed the greatest reduction in bias (0.1 g/dl, p<0.001) whereas BMI (0.07 g/dl, p = 0.0178), RBC transfusion (0.06 g/dl, p<0.001), statins (0.04 g/dl, p<0.05) and beta blocker (0.03 g/dl, p = 0.02) showed a slight effect on bias. Intraoperative substitution of volume and other co-medications did not influence the bias significantly.

Conclusion

Many interventions like substitution of fluids, coagulating factors or RBC units rely on the accuracy of laboratory measurement devices. Although BGA Hb testing showed a consistently stable difference to CL, our data confirm that BGA devices are associated with different bias. Therefore, we suggest that hospitals assess their individual bias before implementing BGA as valid and stable supplement to CL. However, based on the finding that bias decreased with smaller Hb values, which in turn are used for transfusion decision, we expect no unnecessary or delayed RBC transfusion, and no major impact on the LIBERAL trial performance.

Introduction

Acute anemia is one of the most important and common complication during and after surgery [1]. While surgical techniques have advanced over the course of time, intraoperative blood loss is still present in major surgery. Even though infusions or coagulating factors are often administrated to compensate massive blood loss, transfusion of allogenic red blood cells (RBC) is in many cases inevitable [2]. In these critical situations hemoglobin (Hb) and/or hematocrit levels are one key factor among others to determine the need for RBC transfusion. It is noteworthy to mention that transfusion guidelines also recommend to consider physiological constitution, hemodynamic and spirometry parameters, volume status, and dynamic of bleeding.

Several methods, such as central laboratory (CL), blood gas analysis (BGA) or CO-Oximetry [3] are applied to evaluate the Hb level, of which BGA and CL are the most common used methods intraoperatively. However, measurement values differ between BGA and CL [4–7]. Studies compared different devices in order to establish the optimal, less invasive, fastest but at the same time most accurate method to measure Hb level. For example, Giraud et al compared the accuracy of four different bedside devices with CL. The analysis of 219 measurements from 53 patients revealed different Hb levels for each device. Among all methods, the point-of-care testing (POCT) device HemoCue® displayed the smallest and CO-Oximetry the biggest bias [3]. Non-invasive Hb measurement, such as spectrophotometry, turned out to be inferior to invasive Hb measurement (BGA) [8]. Additionally, studies investigated whether differences in clinical routines or patient’s physiological status affect Hb values using different devices for measurements. For example, the concentration of Hb varies depending on the amount of plasma volume, leading to different Hb levels between arterial or venous samples within the same patient. In addition, position of the body or time of the day during blood withdrawal also appear to influence Hb level [9]. In cases of hemodilution with hematocrit below 30%, measurement of Hb concentration using CO-Oximetry seems to be more accurate than conductivity [10]. Ng and colleagues evaluated the bias in Hb level between CL and POCT device before and after major blood loss (≥ 25% of blood volume). The analysis revealed that Hb levels were lower when measured with POCT device as compared to laboratory. Therefore, the authors recommend physicians not to use the POCT device in particular situation such as extensive blood loss when accurate measurement is essential [11]. To the best of our knowledge, however, the bias between BGA measurement and CL depending on the patient’s physiology, co-medication and clinical events such as blood transfusion or intraoperative volume therapy has not been assessed. Additionally, accurate measurement of Hb level is important for studies ultimately implementing clinical guidelines. A number of studies compared surgical outcome in patients receiving either a restrictive or liberal transfusion regime. In these studies, several methods have been used to detect Hb concentration, however details about the used method for measurement are often not described. In addition, it is not evident whether the same methods were used in multicenter studies or more importantly considered during the analysis [12–15].

Here, we performed a sub-analysis of the ongoing “liberal transfusion strategy to prevent mortality and anemia-associated ischemic events in elderly non-cardiac surgical patients” (LIBERAL-Trial) [16] to investigate the accuracy of BGA Hb testing compared to CL by determining factors that potentially influence measurements in surgical patients.

Material and methods

Study design

This sub-analysis was conducted with data from the LIBERAL-Trial (NCT03369210) [16]. The LIBERAL-Trial is a prospective, open, multicenter, randomized phase IV clinical trial to investigate whether a liberal transfusion strategy of RBCs prevents mortality and anemia-associated ischemic events in elderly patients undergoing non-cardiac major surgery. Briefly, elderly patients (≥70 years) scheduled for intermediate or high risk non-cardiac surgery were included in the trial and randomized to a restrictive or liberal transfusion group as soon as Hb level dropped below ≤9 g/dl during surgery or postoperative day 1 to 3. Hemoglobin was measured intraoperatively and on daily basis. The LIBERAL-Trial was approved by the Ethics Committee of the University of Frankfurt (Ref: 139/17F) and by the federal authority (Paul-Ehrlich-Institute) [16].

Blood sampling

During hospital stay, Hb measurement by CL or BGA occurred regularly. Pairs of Hb were taken in the recovery room after surgery. Furthermore, if the patient remained for 24 hours in the recovery room, another pair was taken the following day at 4 am. If patients were admitted to the intensive care unit or intermediate care unit, pairs of Hb were taken at 4am and 4pm on a regular basis. As stated in the study protocol [16] during massive bleeding or before RBC transfusion, additional blood samples were analysed using CL or BGA.

Central laboratory measurements

Hemoglobin level was determined with SYSMEX XN-10® as part of the SYSMEX XN-9000® using the Sodium-Lauryl-Sulfate (SLS) method in the local CL department of all three trial sites. EDTA tubes with either arterial or venous whole blood were used for blood analysis. Analysis with CL was mainly used at the peripheral ward during pre- and postoperative management. Results were displayed approximately within 20 or 60 minutes on weekdays and weekends, respectively.

Blood gas analysis

To determine the Hb level of either arterial or venous heparinized blood samples, the ABL800 Flex® analyzer was used at the LIB-07 center and the GEM Series® at the LIB-26 center. In the LIB-05 center, the RapidPoint 500® was used to measure the Hb level using CO-oximetry. Hemoglobin level was quantified with using the law of Lambert-Beersche. If Hb was <0.16 g/dl (0.1 mmol/l) or >40.26 g/dl (25 mmol/l), results were classified as outlier and not considered in the analysis. Blood samples, both arterial and venous, were taken by trained medical staff and transferred to the BGA analyzer immediately. Results were displayed in the patients electronic medical file within 65 seconds.

Data collection

Blood Gas Analysis and CL measurements are subject to internal and external quality control according to the guidelines of the German Medical Association guidelines in medical laboratory examinations [17]. Data of patients enrolled in the LIBERAL-Trial from January 2018 until October 2019 were extracted and analyzed. Hemoglobin measurements were collected before, during and after surgery until hospital discharge or 30 days postoperative, whichever occurred first [16]. For estimation of intraoperative or postoperative anemia according to the definition of the WHO [18], both Hb measurements through BGA devices and CL were used. We did not discriminate between arterial or venous blood sample as the patient’s electronic case report file does not provide information on sample’s origin. Blood sampling occurred at least every third day. After RBC transfusion, Hb concentration was examined to ensure that the target Hb level had been reached within 24 hours. Several patient characteristics were investigated including Body Mass Index (BMI), gender, 15 most common used co-medication (ACE-Inhibitor (ACE), beta blocker (BETA), calcium channel blockers (CAANT), aspirin (ASS), benzodiazepine (BENZO), statins (STAT), insulin (INS), oral antidiabetic drugs (ODIA), antiarrhythmic agents (AARRH), Parkinson medication (PARK), neuroleptics (NEURO), nonsteroidal anti-inflammatory drugs (NSAR), opioids (OPIO), oral anticoagulants (ACOA) and dual platelet aggregation inhibitor (DPLAT)), volume therapy and smoker status. Volume therapy refers to the administration of crystalloid fluids during surgery. Often, 500mL to 1000mL crystalloid solutions are used. Immediately before surgery, crystalloids are applied to resuscitate the deficit in volume from fasting. During the surgery, crystalloid solutions are applied in order to resuscitate the deficit in blood volume from perspiration or bleeding by using physiological marker such as lactate, blood pressure and heart rate. Blood loss and the amount of crystalloids were not measured. Because volume therapy is part of our perioperative guidelines, we compared Hb pairs after volume therapy with postoperative Hb pairs. In order to investigate whether RBC transfusion affects the bias of Hb level measured with BGA or CL, Hb pairs of patients taken within 24 hours after RBC transfusion were compared with non-transfused patients or with patients in which transfusion occurred more than 24 hours prior to Hb measurement. This timeframe was used because an increase in Hb level is expected within two to 24 hours after RBC transfusion [17].

Statistical analysis

We analyzed pairs of CL and BGA measurements. A pair was defined as two Hb values obtained within 2 hours from CL and BGA. To ensure robust analysis we included only centers with n >200 Hb pairs. Data are presented as mean values with standard deviation. Robust statistical methods such as median absolute deviation (MAD) were used to assess bias for extreme outliers, for example in cases of RBC transfusion or massive blood loss during the two hours time range. Outliners were excluded if bias displayed a value 5 times of standard deviation (SD) of the respective median value. Statistical significance level was accepted with p <0.05. Bland-Altman analysis was performed to analyze the agreement between CL and BGA measurements [19]. LOWESS curve was used to describe the development of bias in decreasing Hb values. Multiple regression analysis and t-test were applied to determine significance of the bias depending on various factors including patient characteristics (BMI, gender, 15 most common used co-medication and smoker status), RBC transfusion and volume substitution. The results are displayed using empirical cumulative distribution functions. Logarithmic scales were used to meet the range of Hb values. Except for center and Hb-level, analysis of several other factors is strictly explorative. Because we expected no differences to our results, we did not adjust for multiplicity. R-Development Core Team (2008), Version 3.6.1, R Foundation for Statistical Computing, Vienna, Austria) was used for statistical analysis.

Results

In total, three centers including 963 patients with 1,814 pairs of Hb measurements were analyzed. The mean difference in Hb pairs of CL and BGA measurements taken within two hours was assessed to determine the biases for center LIB-07 (n = 1307), center LIB-26 (n = 263) and center LIB-05 (n = 244), respectively. Mean bias was comparable between LIB-07 and LIB-26. Empirical cumulative distribution function showed that approximately 95% of all CL values are smaller than BGA (Fig 1). LIB-07 and LIB-26 revealed a mean bias of -0.38 ± 0.28 and MAD ± 0.15 g/dl (median -0.4 g/dl) (Fig 2A) whereas LIB-05 showed a smaller bias (-0.09 ± 0.45 g/dl) but greater MAD (± 0.45 g/dl) (median -0.1) (Fig 2B).

Fig 1. Comparison of the bias.

Fig 1

Open in a new tab

Empirical cumulative distribution function was applied to detect potential bias between centers (LIB-05, LIB-07 and LIB-26). Hb = hemoglobin, CL central lab, BGA = blood gas analysis, LIB-05 = Bonn, LIB-07 = Lrankfurt, LIB-26 = Wuerzburg.

Fig 2.

Fig 2

Open in a new tab

A and B: Histogram with density for hemoglobin pairs from LIB 07 and LIB-26. The relative frequency of Hb difference is displayed (a, b). Sd (Standard deviation), MAD.

In order to avoid interference with different standard deviations from the trial sites that may be caused by the different analytic devices, we focused on two centers using the same BGA technique: LIB-07 and LIB-26. In total 309 patients and 1,570 Hb pairs were included in the further analysis. Overall, 1,389 Hb pairs were measured ≥8 g/dl and 184 Hb pairs between 6 and 8 g/dl. To assess a relation between the difference in each pair of values and the respective mean Hb value, data pairs were plotted using a Bland-Altman Plot (Fig 3). The LOWESS curve for absolute Hb values showed that low Hb values are associated with smaller bias. Hb values of 10 to 15 g/dl display a bias with -0.5 g/dl whereas Hb values of 6 g/dl display a bias of -0.25 g/dl (Fig 3). However, Hb differences plotted as percentage revealed an increase of up to 5% of the Hb value that are associated with smaller Hb values (Fig 4).

Fig 3. Bland-Altman Plot and LOWESS curve showing the differences of the Hb differences (LIB-07 and LIB-26).

Fig 3

Open in a new tab

The difference of the Hb values (Hb-CL and Hb-BGA) calculated for each pair and plotted against the mean value of both measurements are displayed ((Hb-CL + Hb-BGA)/2). The LOWESS curve (red line) shows the tendency of mean bias with lower Hb levels. Hb = hemoglobin, CL = central lab, BGA = blood gas analysis.

Fig 4. Bland-Altman Plot and LOWESS curve showing differences in percentage of Hb pairs (LIB-07 and LIB-26).

Fig 4

Open in a new tab

The difference in percentage of Hb values (Hb-CL and Hb- BGA) calculated for each pair and plotted against the mean value of both measurements are displayed ((Hb-CL + Hb-BGA)/2). The LOWESS curve (red line) shows the tendency of mean bias expressed as percentage of CL Hb with lower Hb levels. Hb = hemoglobin, CL = central lab, BGA = blood gas analysis.

Impact of various patient characteristics on bias

We compared Hb pairs to investigate whether various patient characteristics influence the bias between BGA and CL measurement. Age, gender or volume therapy during surgery did not significantly influence bias (Table 1). Significant differences in mean bias were detected for BMI, smoker status and for patients with and without transfusion. Mean bias was reduced by 0.07 g/dl in underweight (-0.29 ± 0.33 g/dl, BMI < 19) compared to overweight patients (-0.36 ± 0.28 g/dl, BMI > 26) (p = 0.0178) (Table 1) and reduced by 0.1 g/dl in smokers (-0.30 ± 0.24 g/dl) compared to non-smoker (-0.40 ± 0.29 g/dl) (p<0.001). Mean bias between ex-smoker (-0.37 ± 0.28 g/dl) and non-smoker was similar (-0.40 ± 0.29 g/dl) (Table 1, Fig 5). Among 309 patients, 141 received at least transfusion of one unit of RBC and/or autologous blood. Mean bias was reduced by 0.06 g/dl in patients after RBC transfusion (-0.34 ± 0.26 g/dl) compared to patients without transfusion (-0.4 ± 0.29 g/dl) (p<0.001) (Table 1).

Table 1. Possible confounding factors N = absolute number of patients, SD = standard deviation (LIB-07 and LIB-26).

Variable Description N Mean Bias (g/dl) SD Bias (g/dl) P-value
Age < 80 years 224 -0.38 0.29 0.952
> 80 years 85 -0.38 0.27
Gender Male 196 -0.38 0.27 0.935
Female 113 -0.38 0.29
BMI Underweight 11 -0.29 0.33 0.0178
Normalweight 132 -0.37 0.27
Overweight 104 -0.36 0.28
Obesity class I 46 -0.36 0.27
Obesity class II/III 16 -0.45 0.32
Smoker status Smoker 33 -0.30 0.24 0.000016
Ex-smoker 135 -0.37 0.28
Non-smoker 141 -0.40 0.29
Volume therapy Intraoperative 271 -0.39 0.26 0.202
Postoperative 13 -0.37 0.29
After RBC transfusion No transfusion last 24h 168 -0.40 0.29 0.000869
After transfusion 141 -0.34 0.26
Open in a new tab

Fig 5. Bias depending on smoker status (LIB-07 and LIB-26).

Fig 5

Open in a new tab

The relative frequencies of Hb pairs in smoker, non-smoker and ex-smoker are displayed. Hb = hemoglobin, CL = central lab, BGA = blood gas analysis.

Intraoperative volume therapy versus postoperative values

To elucidate whether intraoperative volume substitution with crystalloid fluids affects the bias, we compared blood samples obtained immediately after surgery in the recovery room (n = 271) with samples taken up to 24 hours after skin incision (postoperative, n = 13). Mean bias was non-significantly reduced by 0.02 g/dl in patients after 24 hours of volume substitution (-0.37 ± 0.29 g/dl) compared to values immediately after surgery (-0.39 ± 0.26 g/dl) (p = 0.202) (Table 1).

Impact of co-medications on bias

Several co-medications (ACE, BETA, CAANT, ASS, BENZO, STAT, INS, ODIA, AARRH, PARK, NEURO, NSAR, OPIO, ACOA, and DPLAT) where taken by the patients at hospital admission. BETA and STAT influenced the bias in Hb level significantly (Table 2). Mean bias was reduced by 0.03 g/dl for patients with betablockers (-0.4 ± 0.28 g/dl) compared to patients without (-0.37 ± 0.28 g/dl) betablockers (BETA) (p = 0.02) and by 0.04 g/dl for patients with statins (-0.41 ± 0.28 g/dl) compared to patients without (-0.37 ± 0.28 g/dl) statins (STAT) (p<0.05) (Table 2).

Table 2. Biases depending on co-medication (LIB-07 and LIB-26).

Medication N Mean Bias (g/dl) SD Bias (g/dl) P-Value
AARRH no 298 -0.38 0.28 0.689
yes 11 -0.37 0.25
ACE no 119 -0.37 0.29 0,16
yes 190 -0.39 0.28
ACOA no 240 -0.38 0.28 0.319
yes 69 -0.39 0.29
ASS no 196 -0.38 0.29 0.453
yes 113 -0.39 0.27
BENZO no 303 -0.38 0.28 0.691
yes 6 -0.41 0.31
BETA no 151 -0.37 0.28 0.0234
yes 158 -0.40 0.28
CAANT no 227 -0.38 0.28 0.733
yes 82 -0.39 0.29
DPLAT no 269 -0.38 0.28 0.85
yes 13 -0.38 0.26
INS no 277 -0.38 0.28 0.485
yes 32 -0.36 0.32
Neuro no 295 -0.38 0.28 0.823
yes 14 -0.37 0.29
NSAR no 287 -0.38 0.28 0.704
yes 22 -0.37 0.27
ODIA no 262 -0.38 0.28 0.762
yes 47 -0.39 0.31
OPIO no 267 -0.38 0.29 0.44
yes 42 -0.37 0.22
PARK no 301 -0.38 0.28 0.299
yes 8 -0.43 0.25
STAT no 168 -0.37 0.28 0.00762
yes 141 -0.41 0.28
Open in a new tab

ACE = ACE-Inhibitor, BETA = beta blocker, CAANT = calcium channel blockers, ASS = aspirin, BENZO = benzodiazepine, STAT = statins, INS = insulin, ODIA = oral antidiabetic drugs, AARRH = antiarrhythmic agents, PARK = Parkinson medication, NEURO = neuroleptics, NSAR = nonsteroidal anti-inflammatory drugs, OPIO = opioids, ACOA = oral anticoagulants, DPLAT = dual platelet aggregation inhibitor, N = absolute number of patients, SD = standard deviation.

Discussion

Measurement of diagnostic parameters using BGA is an integral part of clinical routine providing fast results and enabling rapid response in critical situations. However, results of BGA devices need to be regarded cautiously due to conflicting results from various studies. Many interventions like substitution of fluids or coagulating factors rely on the accuracy of laboratory measurement devices [20]. When looking at clinically relevant decisions, for example within randomized trials comparing outcome of different transfusion strategies, accurate measurement of Hb values should be ensured anytime [13–15]. Here, we investigated the accuracy of BGA compared to CL by determining factors that can potentially influence these measurements. In total, 1,814 pairs of Hb measurements of 963 patients of the LIBERAL-Trial have been investigated of which 1,570 pairs of 309 patients were analyzed in detail. Interestingly, the greatest influence on the bias results from different BGA devices used by the trial sites. Mean bias is comparable between ABL800 Flex® analyzer and the GEM series® (-0.38 ± 0.28 g/dl) whereas RapidPoint 500® showed smaller bias but greater MAD. Overall, the Bland-Altman analysis and LOWESS curve revealed that bias decreased with smaller Hb values. We also examined several factors that could lead to bias in Hb values. Of all investigated factors smoker status showed the greatest effect whereas BMI, RBC transfusion, BETA and STAT showed only a slight effect on bias. Bias was reduced by 0.1 g/dl in smokers compared to non-smoker. This might be caused by the different methods used to measure Hb concentration. The SLS-method used by CL is sensitive to methemoglobin, which is increasingly found in smokers and may result in falsely higher concentration of Hb measured by CL [21]. Furthermore, we observed that transfusion of RBC slightly reduced the difference in bias by 0.06 g/dl. Similarly, we found slight differences of 0.07 g/dl for BMI (underweight vs overweight), 0.03 g/dl for BETA and 0.04 g/dl for STAT. It is noteworthy, that intraoperative substitution of volume and co-medications, except BETA and STAT, did not influence the bias significantly. Based on the stable bias with small standard deviation which were largely unaffected by various physiological factors both BGA devices—ABL800 Flex® analyzer and the GEM series®- ensure comparable and accurate estimation of Hb levels.

Our analysis confirms that BGA produces stable values for Hb and is not influenced by the patient’s physiology except for smoker. This observation is especially relevant during trauma setting where no information about the critical patient is available. The bias of smoker is only minor and therefore, we believe, negligible in this setting Hemoglobin measurement is not only important for categorizing the patients’ trauma but is also used in several scoring systems, such as the Trauma Associated Severe Hemorrhage Score (TASH) [22]. Prior to RBC transfusion, the physician evaluates if there are possible transfusion triggers such as low Hb values, hypotension, tachycardia or lactatemia. In these settings, particularly in intensive care units where Hb is frequently determined [23,24], measurements with BGA is recommended since less blood volume is required, turn-around time is faster, and stable results for Hb can be observed. Blood sparing methods consequently lead to fewer RBC transfusions and shorter duration of stay in the intensive care unit [23]. Unfortunately, noninvasive methods tended to be inferior to invasive BGA [8,24]. A, faster turn-around time consequently leads to faster decision making. However, once the laboratory results become available, the treatment should be checked and if necessary adjusted accordingly.

In our analysis, we used data of a large prospective multicenter trial where patients are randomized to a restrictive or liberal transfusion group with a target range for post-transfusion Hb concentration of 7.5–9 or 9–10.5 g/dl, respectively [16]. Our analysis show that BGA is a stable supplement to CL to measure Hb values. In this sub-analysis of the LIBERAL trial, we found that Hb values measured by BGA vary between different centers. To allow a precise analysis each participating center should determine their individual correction factor which will be applied during our final analysis. This study illustrates the importance of determining the concordance between values obtained by BGA and those obtained in the CL for each individual hospital.

One limitation of our analysis is that we were not able to compare pairs of one blood sample. This may result in a different bias due to dynamic fluid and blood shifts, such as massive blood loss or substitution of crystalloid fluids. However, we addressed this possibility in our analysis and excluded outliners by using robust statistical methods: bias deviated only slightly from the median (MAD ± 0.15 g/dl) and thus can be regarded as stable. Here we compared the ABL800 Flex® analyzer, GEM series® and RapidPoint 500®, therefore, our results do not apply to other measurement methods used within clinical routine. Another limitation is the multiple testing we used for this analysis. Ideally, equal number of Hb pairs for the different ranges of ≤6, 6–8, and ≥8 g/dl would be preferred for analysis. However, our observed distribution of higher amount of Hb pairs for 8 g/dl which are followed by 6–8 g/dl represent the clinical distribution of Hb values of patients undergoing non-emergent major surgical procedures. Furthermore, we did not analyze the bias of pre- or intraoperative Hb pairs, which could have contributed to the evaluation of bias during massive bleeding. Finally, due to the study protocol we were not able to differentiate between arterial or venous samplings which may also affect the bias due to possible differences in plasma volume. Since BGA is frequently used in clinical practice to assess Hb status and potentially trigger clinical decisions, we suggest that future trials should consistently use either venous or arterial blood for analysis.

Taken together, we investigated the accuracy of three BGA devices to measure Hb concentration with CL as reference method within the ongoing multicenter randomized controlled LIBERAL-Trial. Our analysis revealed that BGA devices used within the trial are associated with different biases. Of all investigated possible confounders only smoker status was systematically related to bias. Multicenter trials, such as the LIBERAL-Trial, should assess whether transfusion decision was based on Hb concentration estimated with BGA or CL. Our analysis showed that bias increases relatively with lower Hb values. However, clinically these findings are minimal (0.25 g/dl for Hb pairs from 6 to 8 g/dl). This Hb range is used for transfusion decision and we expect no unnecessary or delayed RBC transfusion, and no major impact on the LIBERAL-Trial performance. A small bias is negligible, and clinicians should not hesitate to trust measurements using BGA devices. Nevertheless, we suggest that hospitals assess their individual bias before implementing BGA as valid and stable supplement to CL.

Data Availability

Access restrictions apply to the data underlying the findings. Ethical restrictions prevent public sharing of data since the LIBERAL Trial still is a recruiting randomized controlled trial and the database has not been finalized. Thus, it is protected against unauthorized access. Access can be granted to authorized persons by the sponsor. Data can be obtained by contacting the Clinical Trial Centre Leipzig, University of Leipzig, Härtelstraße 16-19, 04107 Leipzig (info@zks.uni-leipzig.de).

Funding Statement

This work is supported by the German Research Foundation (Deutsche Forschungsgemeinschaft; grant no. ME 3559/3-1 to PM) who had no impact in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript. https://www.dfg.de/en/index.jsp.

References

  • 1.Platz J. and Hyman N., Tracking intraoperative complications. J Am Coll Surg, 2012. 215(4): p. 519–23. 10.1016/j.jamcollsurg.2012.06.001 [DOI] [PubMed] [Google Scholar]
  • 2.Querschnitts-Leitlinien zur Therapie mit Blutkomponenten und Plasmaderivaten. Deutscher Ärzte-Verlag, 2014(4. überarbeitete Auflage).
  • 3.Giraud B., et al. , Comparison of haemoglobin measurement methods in the operating theatre. Br J Anaesth, 2013. 111(6): p. 946–54. 10.1093/bja/aet252 [DOI] [PubMed] [Google Scholar]
  • 4.Dolscheid-Pommerich R.C., et al. , Comparability of Point-of-Care versus Central Laboratory Hemoglobin Determination in Emergency Patients at a Supra-Maximal Care Hospital. PLoS One, 2016. 11(11): p. e0166521 10.1371/journal.pone.0166521 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Rivas Chirino L., et al. , Use of a blood gas analyzer to measure blood hemoglobin during liver transplantation: a study of 935 paired samples. Transplant Proc, 2006. 38(3): p. 810–1. 10.1016/j.transproceed.2006.01.051 [DOI] [PubMed] [Google Scholar]
  • 6.Gavala A. and Myrianthefs P., Comparison of point-of-care versus central laboratory measurement of hematocrit, hemoglobin, and electrolyte concentrations. Heart Lung, 2017. 46(4): p. 246–250. 10.1016/j.hrtlng.2017.04.003 [DOI] [PubMed] [Google Scholar]
  • 7.Miller R.D., et al. , A comparison of three methods of hemoglobin monitoring in patients undergoing spine surgery. Anesth Analg, 2011. 112(4): p. 858–63. 10.1213/ANE.0b013e31820eecd1 [DOI] [PubMed] [Google Scholar]
  • 8.Lamhaut L., et al. , Comparison of the accuracy of noninvasive hemoglobin monitoring by spectrophotometry (SpHb) and HemoCue(R) with automated laboratory hemoglobin measurement. Anesthesiology, 2011. 115(3): p. 548–54. 10.1097/ALN.0b013e3182270c22 [DOI] [PubMed] [Google Scholar]
  • 9.Berkow L., Factors affecting hemoglobin measurement. Journal of Clinical Monitoring and Computing, 2013. 27(5): p. 499–508. 10.1007/s10877-013-9456-3 [DOI] [PubMed] [Google Scholar]
  • 10.Myers G.J. and Browne J., Point of care hematocrit and hemoglobin in cardiac surgery: a review. Perfusion, 2007. 22(3): p. 179–83. 10.1177/0267659107080826 [DOI] [PubMed] [Google Scholar]
  • 11.Ng W.L., et al. , Accuracy and reliability of the i-STAT point-of-care device for the determination of haemoglobin concentration before and after major blood loss. Anaesth Intensive Care, 2014. 42(4): p. 495–9. 10.1177/0310057X1404200411 [DOI] [PubMed] [Google Scholar]
  • 12.Villanueva C., Gastrointestinal bleeding: Blood transfusion for acute upper gastrointestinal bleeding. Nat Rev Gastroenterol Hepatol, 2015. 12(8): p. 432–4. 10.1038/nrgastro.2015.116 [DOI] [PubMed] [Google Scholar]
  • 13.Carson J.L., et al. , Liberal or restrictive transfusion in high-risk patients after hip surgery. N Engl J Med, 2011. 365(26): p. 2453–62. 10.1056/NEJMoa1012452 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Mazer C.D., et al. , Restrictive or Liberal Red-Cell Transfusion for Cardiac Surgery. N Engl J Med, 2017. 377(22): p. 2133–2144. 10.1056/NEJMoa1711818 [DOI] [PubMed] [Google Scholar]
  • 15.Holst L.B., et al. , Lower versus higher hemoglobin threshold for transfusion in septic shock. N Engl J Med, 2014. 371(15): p. 1381–91. 10.1056/NEJMoa1406617 [DOI] [PubMed] [Google Scholar]
  • 16.Meybohm P., et al. , Liberal transfusion strategy to prevent mortality and anaemia-associated, ischaemic events in elderly non-cardiac surgical patients—the study design of the LIBERAL-Trial. Trials, 2019. 20(1): p. 101 10.1186/s13063-019-3200-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Richtlinie der Bundesärztekammer zur Qualitätssicherung laboratoriumsmedizinischer Untersuchungen—Rili-BÄK. Deutsches Ärzteblatt, 2014. Jg. 111(38). [Google Scholar]
  • 18.WHO., Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. Vitamin and Mineral Nutrition Information System. Geneva, World Health Organization, 2011. WHO/NMH/NHD/MNM/11.1. [Google Scholar]
  • 19.Bland J.M. and Altman D.G., Measuring agreement in method comparison studies. Stat Methods Med Res, 1999. 8(2): p. 135–60. 10.1177/096228029900800204 [DOI] [PubMed] [Google Scholar]
  • 20.Whiting P., et al. , Viscoelastic point-of-care testing to assist with the diagnosis, management and monitoring of haemostasis: a systematic review and cost-effectiveness analysis. Health Technol Assess, 2015. 19(58): p. 1–228, v-vi. 10.3310/hta19580 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Dörner K., Klinische Chemie und Hämatologie. Georg Thieme Verlag, 2006(6. Auflage): p. 250. [Google Scholar]
  • 22.Maegele M., Acute traumatic coagulopathy: Incidence, risk stratification and therapeutic options. World J Emerg Med, 2010. 1(1): p. 12–21. [PMC free article] [PubMed] [Google Scholar]
  • 23.Riessen R., et al. , A Simple "Blood-Saving Bundle" Reduces Diagnostic Blood Loss and the Transfusion Rate in Mechanically Ventilated Patients. PLoS One, 2015. 10(9): p. e0138879 10.1371/journal.pone.0138879 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Causey M.W., et al. , Validation of noninvasive hemoglobin measurements using the Masimo Radical-7 SpHb Station. Am J Surg, 2011. 201(5): p. 592–8. 10.1016/j.amjsurg.2011.01.020 [DOI] [PubMed] [Google Scholar]

Decision Letter 0

Laura Pasin

Transfer Alert

This paper was transferred from another journal. As a result, its full editorial history (including decision letters, peer reviews and author responses) may not be present.

24 Jul 2020

PONE-D-20-13621

Factors Influencing the Bias between Blood Gas Analysis versus Central Laboratory Hemoglobin Testing. A secondary analysis of a randomized controlled trial.

PLOS ONE

Dear Dr. Meybohm,

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.

As pointed out by all reviewers, the topic is interesting, but the study suffers from several limitations. In particular, I strongly believe that professional statistical assistance should be considered. Moreover, clinical implications of your results should be discussed in depth. 

Please ensure that your decision is justified on PLOS ONE’s publication criteria and not, for example, on novelty or perceived impact.

Please submit your revised manuscript by 23/08/20. 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

We look forward to receiving your revised manuscript.

Kind regards,

Laura Pasin

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. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[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: Partly

Reviewer #2: No

Reviewer #3: Partly

Reviewer #4: Partly

**********

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

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: Yes

Reviewer #4: No

**********

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: No

Reviewer #3: Yes

Reviewer #4: 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

Reviewer #3: Yes

Reviewer #4: 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: • In the introduction, the authors stated: “hemoglobin (Hb) and/or hematocrit levels are one key factor to determine the need for RBC transfusion”. However, many studies have already demonstrated that hemoglobin concentration is not a reliable indicator of transfusion requirement. The authors should add a comment about that.

• While in the abstract the objective of the study is “to investigate the accuracy of BGA hemoglobin testing as compared to CL services”, in the main manuscript the declared aim is different, particularly “to elucidate factors influencing the bias between the most common used methods -CL and BGA- to measure Hb levels in surgical patients.” Which one is the right aim?

• In the study design section, the authors stated that elderly patients scheduled for surgery were included in the trial, but they do not clarify the lower limit of age for inclusion in the study.

• In the result section, the authors stated: “mean bias was reduced by

0.02 g/dl in patients after 24 hours of volume substitution (-0.37 ± 0.29 g/dl) compared to values immediately after surgery (-0.39 ± 0.26 g/dl)”. However, p value is not statistically significant (p=0.202).

• The authors did not discuss the impact of their results on the clinical practice. How could a physician use these results in his daily work when deciding whether to transfuse a patient?

Reviewer #2: The authors set out to study Bias between CL and three different BGA machines. One of them was Co-oximeter. Ideally the study design for this research question would be significantly different from what the authors have used. The design would involve a single blood sample taken per patient and sent for CL and BGA analysis, with stringent control of analytical factors related to specimen collection, transport and processing.

The range of Hb, which is of clinical interest – i.e. where the decisions regarding transfusion are usually made (6 to 8 gm/dl), is poorly represented. Ideally the ranges of Hb should be equally represented <6, 6-8, >8…for example.

I don’t agree with Authors claiming that the bias decreases with lower Hb values. Firstly, there is no explanation provided that why this might be true and secondly, as mentioned before the low range is poorly represented.

The authors then go on to analyse the factors that impact the bias. They have included several factors (patient related and treatment related), without explanation or rationale as to why they included these factors. This exercise then appears to be similar to ‘data mining’ exercise rather than carefully thought of hypothesis generation and testing.

Authors have not taken into consideration the fact that one patient is represented in the data more than once. This fact is important when analysing patient related factors. This is also important as the three centres were very different. For example the Lib07 centre had 6 times more specimens than the other centres. However, the number of patients of the Lib07 and Lib26 combined were only 1/3rd of the entire sample size (result section 2nd Para). This means the number of samples per patient were much more in Lib 07 than others.

I am concerned that the authors have found incidental findings in a poor study design, which are further poorly substantiated in the discussion section. The Authors do talk about smokers and MetHb. The Cooximeter should be able to pick this up in contrast to other BGL machines assessed. So did this effect the bias in the Lib05 centre?

The Authors have used regression analysis to study several factors. Apart from above criticism of lack of rationale/procedure in selecting factors, there is a lot unknown regarding collinearity of factors, excessive representation of one patient etc. Was the assumption of normality met?

The authors have discussed about bias, but the biases appear to be minimal which makes you wonder if its all clinically meaningful. Also authors have used statements such as “Mean bias was reduced by 0.02 g/dl in pts after 24 hours of volume substitution compared to values after surgery” (In results section). Basically there is no difference given the P value and insignificant difference. Hence commenting as the bias is reduced, is incorrect.

The authors took “results within two hours” to define a pair. The actual time of sampling could be 20 minutes to 60 minutes prior to results.

How did they figure out that there was an intervention between the sampling for BGA and CL? How did they manage these pairs where there was an intervention?

We cant compare venous and arterial blood. Authors do acknowledge that in their limitations. I am surprised by their use of word “might”. It definitely affects the bias. We don’t know what was the extent of this confounding factor.

Overall this analysis should be a small part of the Liberal trial results, rather than a full original text publication.

Reviewer #3: In this manuscript, dr Meybohm and colleagues present results of a secondary analysis of an ongoing mRCT on management of blood transfusion after surgery. The aim of their study is to compare Hb levels measured with BGA to Hb levels obtained through central laboratory analysis. They identified several factors that could potentially bias the results of HB analysis by BGA versus CL analysis.

As BGA analysis is frequently used in clinical practice to assess Hb status and potentially trigger clinical decisions, the results of this study may be potentially relevant for several clinicians.

I have a few comments for the Authors which I hope will help them to improve their manuscript:

1. In my opinion, the Authors should better detail in the Discussion the clinical relevance of their findings, and how the results of their study may affect clinical practice and the ongoing LIBERAl trial. E.g. is a bias of 0.09 g/dL clinically relevant?

2. The Authors should better underline in the introduction the difference between their work and previously published studies on difference between BGA and CL analysis. What does this study add to current knowlegde?

3. I understand that blood samples for BGA and CL analysis were taken at different time-points. However, this is not entirely clear from the Methods. I suggest to describe the blood sampling timing in a specific paragraph "blood sampling"

4. Do the Authors have data on BGA and CL agreement during massive transfusion/blood loss? This would be a interesting subgroup analysis

5. The Authors frequently refer to "volume therapy". However, it is unclear what does this mean and how was assessed. Do the Authors refer to maintenance fluids, volume resuscitation, fluid balance? How were these assessed/managed?

6. Please move the list of medications included in the analysis in the Methods section rather than in the Results section

7. The sentence "The LOWESS curve showed a tendency of the bias with decreasing Hb levels." is unclear, please edit.

8. Do the Authors have data on bias of BGA vs CL assessed before, during, and after surgery?

Reviewer #4: This is an interesting subanalysis of the LIBERAL trial, and the restriction to centres with a large number of cases is a sensible one here in the first instance - one does of course want to know whether case volume actually affects accuracy and this could be usefully explored.

What we have here is a number of paired data points by 2 different methods, but also >1 point per patient. So there is clustering here as well and the pairs are not independent. The statistical methodology needs to allow for this but nowhere is this multiplicity of measurements and how it is handled described. This is particularly important in the case of the predictors of bias.

In Figure 3 it would be interesting to look at percentage bias as well here just to see if the trend here actually represents a different type of measurement error. The issue of whether variance depends on value is also interesting and could usefully be explored.

**********

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: Yes: Shivesh Prakash

Reviewer #3: No

Reviewer #4: 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.

PLoS One. 2020 Oct 30;15(10):e0240721. doi: 10.1371/journal.pone.0240721.r002

Author response to Decision Letter 0

20 Aug 2020

All comments have been addressed in the Rebuttal Letter. We thank the editor and reviewers for their helpful remarks.

Attachment

Submitted filename: Rebuttal Letter_Tanner et al_2020.08.16 .docx

Click here for additional data file. (6.6MB, docx)

Decision Letter 1

Laura Pasin

2 Oct 2020

Factors Influencing the Bias between Blood Gas Analysis versus Central Laboratory Hemoglobin Testing. A secondary analysis of a randomized controlled trial.

PONE-D-20-13621R1

Dear Dr. Meybohm,

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,

Laura Pasin

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 #4: 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 #4: (No Response)

**********

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

Reviewer #1: Yes

Reviewer #4: (No Response)

**********

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 #4: (No Response)

**********

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 #4: (No Response)

**********

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: The authors have well responded to the requests.

The findings are presented in a readable way with nuanced conclusions way, and the limitations are well addressed.

I have no further comments.

Reviewer #4: (No Response)

**********

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: No

Reviewer #4: No

Acceptance letter

Laura Pasin

12 Oct 2020

PONE-D-20-13621R1

Factors Influencing the Bias between Blood Gas Analysis versus Central Laboratory Hemoglobin Testing. A secondary analysis of a randomized controlled trial.

Dear Dr. Meybohm:

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. Laura Pasin

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Attachment

    Submitted filename: Rebuttal Letter_Tanner et al_2020.08.16 .docx

    Click here for additional data file. (6.6MB, docx)

    Data Availability Statement

    Access restrictions apply to the data underlying the findings. Ethical restrictions prevent public sharing of data since the LIBERAL Trial still is a recruiting randomized controlled trial and the database has not been finalized. Thus, it is protected against unauthorized access. Access can be granted to authorized persons by the sponsor. Data can be obtained by contacting the Clinical Trial Centre Leipzig, University of Leipzig, Härtelstraße 16-19, 04107 Leipzig (info@zks.uni-leipzig.de).

    Articles from PLoS ONE are provided here courtesy of PLOS

    RESOURCES