WO2020193839A2 - Method and device for the visual detection of the degree of suitability of a sample isolated from blood for its validation in clinical analyses - Google Patents

Abstract

The present invention relates to a method for visually identifying the degree of suitability of a sample isolated from blood for the validation of its subsequent clinical analysis, taking into consideration the degree of hemolysis of the sample. The method includes the comparison of the plasma of the sample with a degraded, continuous scale divided into seven stages that are distributed, with an equal size, from 55° to 00° and from 00° to 351°, 00° being red in the colour wheel of the double-cone model. The invention also relates to the degraded scale and to a device containing same, which can be made of different materials (for example, paper, card, foam board, plastic or metal).

Description

PROCEDURE AND DEVICE FOR THE VISUAL DETECTION OF THE DEGREE OF FITNESS OF AN ISOLATED BLOOD SAMPLE FOR ITS VALIDATION

IN CLINICAL ANALYSIS

TECHNICAL SECTOR

The present invention falls within the field of clinical analysis, more specifically, in the identification of the degree of hemolysis exhibited by isolated blood samples.

BACKGROUND OF THE INVENTION

Hemolysis is a phenomenon that can alter the results of analytical parameters. If no cross-interference occurs, generally, hemolysis increases the concentration of those determinations that present high concentrations at the intracellular level, and decreases those that are found in low concentrations. For example, Italo Moisés Saldaña collects a study of the interference of hemolysis on 25 biochemical constituents, of which 16 presented clinically significant interference (Saldaña, I.M. An Fac Med 2015; 76 (4): 377-384).

To determine the degree of hemolysis of an isolated blood sample before proceeding to its clinical analysis, there are several methods. Fundamentally, methods are used based on the determination of the absorption spectrum of the sample by spectrophotometry, or visual methods. The most frequently used detection method is visual inspection, however, it is a method that, in general, is considered unreliable (Gómez Rioja, R., et al. Rev Lab Clin. 2009; 2 (4): 185-195). On the other hand, methodologies that focus on sample absorbances are more sensitive, but their application also has drawbacks: it involves additional costs and relatively long time requirements.

Various authors present visual detection hemolysis scales through colorimetry or photographs of eppendorf tubes to show the intensity of hemolysis (Saldaña, IM An Fac Med 2015; 76 (4): 377-384; Plumhoff EA, et al. Mayo Medical Laboratories Communiqué 2008; 33 (12): 1-8). Other authors bet on the relevance of spectrophotometric techniques for the detection of hemolysis in isolated blood samples for having greater precision (Appierto, V. et ai. Bioanalysis. 2014; 6: 1215-1226). There are also comparative studies of other methods currently available, such as serum microRNA detection (Shash, JS et al. PLoS ONE 2016; 1 1 (4) 1-12); In this work, the sensitivity of four different methods for the detection of low levels of hemolysis is analyzed and it is concluded that, of those analyzed, only the detection of microRNA and spectrophotometry are sufficiently sensitive.

On the other hand, different cross interferences have been described in the determination of the hemolysis of a sample that could bias the identification of its intensity, both when using spectrophotometric methods and when using visual detection methods in plasma, and therefore therefore, interfere with some analytical determinations. Cross-interference due to bilirubin and lipemia are especially important. Lipemia overestimates and bilirubin underestimates the intensity of hemolysis and the accuracy in the determination of blood constituents (Gómez Rioja, R., et al. Rev Lab Clin. 2009; 2 (4): 185-195). There is a non-linear relationship between the concentration of free hemoglobin in plasma and the determination of total bilirubin, producing an underestimation in small concentrations of hemolysis and vice versa (Saldaña, I.M. An Fac Med 2015; 76 (4): 377-384).

The International Organization for Standardization addresses the requirements of the quality system to be applied in the clinical laboratory (ISO 15189: 2012), focused on patient safety. Venipuncture blood samples are the most common type of biological sample collected and sent to laboratories for analysis, diagnosis, and therapeutic monitoring. According to the action guidelines of the World Health Organization, the laboratory must detect the presence of hemolysis in the isolated sample to analyze and assess whether it is a reason for rejection for any of the requested determinations (World Health Organization. Diagnostic Imaging and Laboratory Technology. (2002). Use of anticoagulants in diagnostic laboratory investigations. Geneva: World Health Organization. Available at: http: / AAAAAA /, wh¾ snt / sris / handje / 10665/65957).

In the absence of a validated method that entails a low cost and short application time, it is still necessary to find a method for determining hemolysis in isolated blood samples, for clinical analysis, that is fast enough, cheap and reliable.

EXPLANATION OF THE INVENTION

Procedure and device for the visual detection of the degree of suitability of an isolated blood sample for its validation in clinical analysis.

One aspect of the present invention relates to a method for identifying the suitability of an isolated blood sample for subsequent clinical analysis thereof. This procedure comprises a step in which the isolated blood sample to be analyzed is compared with a scale degraded according to the visual intensities of hemolysis in plasma, which incorporates the entire range of color intensities from 55 ° to 00 °. and 0 to 351 ° ^or, being 00 ° the red color on the color wheel. These values reflect the number of degrees of rotation around a double colored cone starting from the color red (00 °).

In this specification, a degraded scale is understood to be a colorimetric scale in which the intensity of the color increases progressively, continuously. To carry it out, the tone in degrees, saturation and brightness in a double cone of colors and the HSL model (from the acronyms Hue, Saturation and Lightness - tone, saturation, luminosity -) were used, as well as information from 277 samples of plasma.

The double cone is made up of two cones that sit on the same circular base and whose vertices are aligned with the center of the circle of said base.The upper and lower vertices correspond to the luminosity, white (100%) and black (0 %) respectively; the distance to the axis with saturation, whose maximum intensity of each color is represented by 100% and the minimum intensity, which corresponds to a gray shadow, with 0%; and the angle corresponds to the color tone. Hue resides along the circular base of the double cone, forming a color wheel that is defined by three primary values by their position on the color wheel. The position of the three primary values are normalized, forming a triangle, as follows: primary red is located at 0 ^or, the primary green to 120 ° and the primary blue 240 °, back to red when returned to the origin of the circle at 360 °.

Three other secondary (subtractive) values are formed between the spaces of the first triangle, giving rise to a second triangle: yellow is at 60 °, cyan at 180 °, and magenta at 300 °. Intermediate hues are generated by transitions between adjacent pairs of the positions of the primary and secondary (subtractive) colors, along the color wheel (orange, turquoise, etc.). Hence the position of the colors can be specified in angle degrees (the full circle being 360 °).

The CMYK model (from the acronym Cyan, Magenta, Yellow and Key) is a subtractive color model that allows to represent a wide range of colors and is well adapted to industrial environments. This model is based on the mixture of pigments of cyan, magenta, yellow and black colors to create others and is the one that has been used to convert the colors obtained according to the HSL model to printing values.

Therefore, one aspect of the invention refers to a method for visually identifying the degree of suitability of an isolated blood sample for validation in clinical analysis, by detecting the degree of hemolysis that the sample has suffered, which includes:

- obtaining plasma from the sample,

- comparison of the plasma color of the sample with a color gradient scale that includes the entire range of color intensities from 55 ° to 00 ° and 00 ° to 351 °, with 00 ° being the color red in the chromatic circle of the double cone and in which said range of intensities is divided into 7 intervals of equal size that correspond to the following stages: stage 1: 55 ° -44 °; stage 2: 43 ° -38 °; stage 3: 37 ° -26 °; stage 4: 25 ° -16 °; stage 5: 15 ° -07 °; stage 6: 06 ° -01 °; stage 7: 00 ° -351 °, whose harmonization in saturation and brightness, respectively, are: stage 1: 35% - 44%, 99% - 89%; stage 2: 47% -50%, 90% -90%; stage 3: 50% -60%, 93% - 89%; stage 4: 68% -80%, 95% - 80%; stage 5: 71% -81%, 91% -78%; stage 6: 68% -83%, 80% -

70%; stage 7: 82% -88%, 70% - 55% (Table 1),

- the assignment of the sample to a specific stage of the 7 stages of the degraded scale described in the previous step.

Table 1. Definition of the stages of the degradation scale according to the HSL model Not all blood parameters are affected to the same extent by hemolysis of the blood sample, so the suitability of the sample must be taken into account in each of the determinations made with each sample. In other words, each of the 7 stages of the degraded scale corresponds to an intensity of hemolysis that implies the potential alteration of certain blood parameters, the number of which increases as the degree of hemolysis increases in such a way that the number of alterations are cumulative to as the intensity of hemolysis in plasma EDTA K2 / K3 increases in healthy adults. For this reason, the parameters that have been documented as potentially altered in blood samples with hemolysis, depending on the intensity of hemolysis (mg / dl of deoxyhemoglobin), included in each of the stages of the degraded scale are reflected below. The data are represented as follows: potentially alterable parameters (stage number encompassing all previous analytes); (number of other stages whose parameters are potentially affected).

- Stage 1. Hydroxybutyrate Dehydrogenase (1)

- Stage 2. Aspartate aminotransferase (2); (one).

- Stage 3. Total bilirubin (3); (two); (one).

- Stage 4. Alanine Aminotransferase; pseodocholinesterase not inhibited and inhibited bydracaine; Creatinine kinase; Ethanol; Gamma-Glutamyltrasferase; Iron; Ammonia (4); (3); (two); (one).

- Stage 5: Ferritin; Myoglobin; Phosphate (inorganic); Rheumatoid factors (5); (4); (3); (two); (one).

- Stage 6: Alanine aminotransferase; Complement C4; Total cholesterol; C-reactive protein; Soluble transferrin receptor; Triglycerides; Valproic acid; Vancomycin (6); (5); (4); (3); (two); (one).

- Stage 7: alpha-acid glycoprotein; Alpha 1-Antrypsin; Albumin; Apolipoprotein A-1; Apolipoprotein B; Antistreptolysin O; Carbamazepine; Creatinine; Ultrasensitive C-reactive protein; Digoxin; Gentamicin; HK glucose; HDL cholesterol; Immunoglobulin A; Immunoglobulin G; Immunoglobulin M; Lithium; Phenobarbital; Phenytoin; Salicylate; Theophylline; Total protein; Uric acid; Urea / BUN (7); (6); (5); (4); (3); (two); (one).

In the event that the plasma of an isolated blood sample is the object of analysis of the parameters stated in the previous paragraph, it is visually evaluated its intensity of hemolysis and is assigned to one of the 7 stages. Later, it will be analyzed and the result of each altered parameter can be evaluated in combination with the stage to which it had been assigned. The stage number to which the sample is assigned must accompany the result of the parameters evaluated in the report, as an alert to the person responsible for the validation of results. Thus, the results will be validated or rejected if, not being in a normal range, its corresponding stage of hemolysis reaches a sufficient intensity to be compatible with a potential alteration of the results.

In the event that the plasma from an isolated blood sample is analyzed for additional parameters to those listed in the previous paragraph, its intensity of hemolysis is visually evaluated and assigned to one of the 7 stages. Subsequently, it will be analyzed and the result of each altered parameter can be evaluated in combination with the stage of the degraded scale. The number of the stage with which the sample is identified must accompany the sample result in the report, as an alert to the person responsible for validating the results. Thus, the results will be validated or rejected if, not being within a normal range, its corresponding stage of hemolysis reaches a sufficient intensity to be compatible with a potential alteration of the results of the parameters analyzed. In the same way, if the results of a parameter are not in the normal range and, for that sample, the hemolysis stage is not compatible with an expected alteration of those observed results, the physiological or pathological alteration of the patient is determined. and therefore the need to repeat a sample is eliminated and costs are reduced.

In the event that an isolated blood plasma sample is subject to analysis of analytes of active drug principles, as in clinical trials, the intensity of hemolysis sufficient to trigger an alteration of the analyte, facilitates the choice of final samples to analyze. In such a way that it avoids biases in the results of research on drugs and costs of the analysis of invalid samples.

As a result of the procedure of the invention, for an altered parameter compatible with alteration by hemolysis according to the stage of the degraded scale to which it is assigned, the person in charge of validation of the laboratory can issue the judgment of suspicion of an alteration of the plasma sample and not to a potential pathology of the patient, Or, it can make a judgment of altered results not expected due to a certain intensity of hemolysis in the sample. Therefore, by means of the method of the invention, the person responsible for the clinical analysis can decide whether or not to carry out a second analysis of a second isolated blood sample.

In this procedure, a device can be used in which the color degraded scale preferably occupies a surface 210 mm high by 297 mm long, each stage corresponding to a size of 105 mm high by 41 mm wide. , the set of the scale colors occupies 105 mm high by 287 long, with a margin of 52.5 mm in the upper part as in the lower part and 5 mm in the right part as in the left part in which the color is , preferably, matt white, and it can be manufactured in practically any material: paper, cardboard, foam board, plastic, metal.

A second aspect of the present invention refers to a degraded color scale to identify the degree of suitability of an isolated blood sample for validation in clinical analysis, depending on the degree of hemolysis that said sample presents, which includes the entire range of color intensities from 55 ° to 00 ° and from 00 ° to 351 °, where 00 ° is the red color in the double cone chromatic circle and in which said range of intensities is divided into 7 intervals of equal size corresponding to the following stages: stage 1: 55 ° -44 °; stage 2: 43 ° -38 °; stage 3: 37 ° -26 °; stage 4: 25 ° -16 °; stage 5: 15 ° -07 °; stage 6: 06 ° -01 °; stage 7: 00 ° -351 ° Whose harmonization in saturation and brightness, respectively are: stage 1: 35% - 44%, 99% - 89%; stage 2: 47% -50%, 90% -90%; stage 3: 50% -60%, 93% - 89%; stage 4: 68% -80%, 95% - 80%; stage 5: 71% -81%, 91% -78%; stage 6: 68% -83%, 80% - 70%; stage 7: 82% -88%, 70% - 55%. This degraded scale can be used according to the procedure indicated in the previous paragraph.

On the other hand, to use the color degraded scale to identify the degree of hemolysis of an isolated blood sample, a device can be made in which the scale preferably occupies a surface 210 mm high by 297 mm long, each stage corresponding to a size of 105 mm high by 41 mm wide, the set of scale colors occupies 105 mm high by 287 long, with a margin of 52.5 mm at the top as in the bottom and 5 mm on the right and left in which the color is preferably matte white. East The device can be made of practically any material: paper, cardboard, foam board, plastic, metal.

BRIEF DESCRIPTION OF THE DRAWINGS

To complement the description that is being made and in order to help a better understanding of the characteristics of the invention, figures are attached as an integral part of said description in which, with an illustrative and non-limiting nature, the following has been represented:

Figure 1. Degraded scale for visual detection of hemolysis with indication of the values used for its preparation.

Figure 2. Example of a degraded scale for the visual detection of hemolysis, to be used in the clinical analysis laboratory.

Figure 3. Heteroscedasticity between stages and relationship between absorbance and stages.

PREFERRED EMBODIMENT OF THE INVENTION

The present invention is further illustrated by the following examples, which are not intended to be limiting of its scope.

Example 1. Recruitment of samples.

Recruitment for convenience of plasma samples in healthy adults, blind and single sample for visual detection of colorimetric hemolysis with standardized reference of absorbances.

The sample selection criteria were the following:

- Men and women.

- Age: between 18 and 55 years old.

- Bilirubinemia in ranges of normality. Monitoring through blood tests.

- Prepandial blood samples (*).

- No Smoking.

- Not having a diagnosed disease.

- Have blood parameters within normal ranges. Do not carry infectious diseases.

(*) Some postprandial samples can exceptionally be included if they are taken as late as possible; the most common cause of lipemia is the recent intake of a meal high in saturated fat, so it will be necessary to apply a correction factor to these samples.

Individuals who had a total bilirubinemia between 0.2 and 1.2 mg / dl were included. Those who obtained a total bilirubinemia result greater than 1.5 of the upper limit of normal and obtained a direct bilirubin result equal to or greater than 35% of total bilirubin were excluded both for the elaboration of the scale and for its validation. . The measurement was carried out by means of a safety blood analysis between 7 and 5 days before obtaining the sample for use in the elaboration and validation of this scale.

Individuals who had triglyceride values, fasting for 8-10 hours, less than 150 mg / dl were included. Those who reached these limits or higher results in the fasting state were excluded both for the elaboration of the scale and for its validation. Measurement was performed by a safety blood test between 7 and 5 days before obtaining the sample for use on this scale. The samples used for the elaboration and validation of the degraded scale were obtained in fasting for at least 5 hours.

Example 2. Elaboration of the scale.

An original scale of degraded colorimetry was developed according to the visual intensities of hemolysis in plasma. The information from 277 plasma samples was used to construct a degraded scale, in which the intensity of the color was progressively increased continuously. For this, saturation, luminosity and hue in degrees were used according to the HSL model and its correlation with the CMYK codes.

As can be seen in figure 1, the scale was designed to include from 55 ° to 00 ° and from 00 ° to 351 °, with 00 being the red color in the chromatic circle at the base of the double cone. These values reflect the number of degrees of rotation along the double cone base color circle starting from the color red (00 °). The 277 samples were evaluated by spectrophotometry as indicated in Example 3. The color tones expressed in degrees and the spectrophotometry results are shown in Figure 1: the degrees at the top of the degraded scale and the hemolysis absorbances at the bottom of the scale degrades. In both cases, the values used in the following examples for the study of concordances and in the statistical tests carried out in the validation process of the hemolysis detection method in isolated blood samples have been marked in bold. From top to bottom, figure 1 shows:

- First line: adjusted color tone degrees that delimit each stage.

- Second line: degrees of the color tone found in the experimental field work of the invention.

- Third line: hemolysis absorbance limits found in the 277 samples used.

- Fourth line: arithmetic mean calculated from the hemolysis absorbance limits found in the 277 samples used (third line) from the extremes of each stage.

From the minimum intensity to the maximum intensity of absorbances, the total of the scale was divided into 7 stages, that is, into 7 fragments of equal size that represent 7 phases or successive stages of hemolysis that an isolated sample of blood can present, of less to greater intensity of hemolysis.

In this example, the scale was designed with a size of 105 mm high by 287 mm long, with each stage corresponding to a size of 105 mm high by 41 mm wide. A device was manufactured using 90 gram paper, in matt colors. As can be seen in Figure 1, the 7 stages into which the scale was divided were delimited by the intervals detailed in Table 2:

Table 2. Intervals of the degraded scale in degrees of color tone and in dimensionless units of absorbance in spectrophotometry.

The scale was designed so that all the stages occupy the same space in the total scale, which avoids visual distortion biases.

Among the 7 stages, the clinical relevance has been considered to be between stage 1 and stage 2. A blood sample with stage 1 hemolysis is considered not to have a clinically relevant impact on most blood measurement parameters in plasma. In stage 1, only one parameter is altered, which, furthermore, is not routinely assessed in clinical tests; in stage 2, a parameter that is analyzed more routinely is potentially altered, and in subsequent stages even more values are altered.

Example 3. Baseline hemolysis titration.

To validate the degraded scale elaborated according to example 2, a spectrophotometric methodology was used since, although it is not the most discriminative technique of those currently existing, in general, it is considered sufficient for the detection of hemolysis in isolated blood samples.

A NanoDrop ™ 2000 Spectrophotometer (Thermo Fisher Scientific Inc., Wilmington, United States of America) was used, with a broad ultraviolet-visible spectrum (190-840 nm) and requiring microvolumes of 0.5 to 2 µl of initial sample. Using this spectrophotometer, highly concentrated samples can be measured without the need to dilute them beforehand.

In this example, in each sample, the absorbance of deoxyhemoglobin was measured across the spectrum through 2 µl of plasma. The plasma of each sample was obtained as indicated in example 4. The NanoDrop ™ 2000 Spectrophotometer was calibrated with 2 ml of distilled water, the measurement was carried out and, once the dimensionless unit of absorbance was obtained in the NanoDrop 2000 / 2000c software 1.6.198 Thermo Fisher Scientific Inc., cleaned with disposable tissues. Next, 2 ml of plasma was taken from an isolated blood sample and analyzed, repeating the process at least 3 times, wiping with a tissue between each measurement. Between the third measurement of a sample and the first measurement of the next sample, the zone of the NanoDrop ™ 2000 Spectrophotometer that comes into contact with the sample was cleaned with 2 µl of distilled water and 1 or 2 measurements were made to corroborate that there was no carryover from the analysis of one sample to the next.

Example 4. Validation procedure of the degraded scale.

In each visual observation (test and retest) there were two observers. The two observers in the test of a specific batch of samples were the same as the two observers in the retest of that batch. The observers were established based on:

- Title,

- experience in hemolysis inspection and sample processing in the laboratory,

- availability of collaboration in the laboratory at three times each day.

The observers received a 10-minute pre-training to perform the evaluations. Subsequently, they were allowed to take a maximum of one minute for the evaluation of each sample.

From each isolated blood sample, blood plasma was obtained. For this, the samples were collected in EDTA (ethylenediaminetetraacetic acid) tubes since it is the most used as an anticoagulant because it interferes with very few blood determinations. Plasma was obtained by centrifuging each sample at 3400 rpm for 10 minutes at 4 ° C, in a maximum of two hours after sampling. The plasma content was transferred to the largest possible number of 1.4 ml tubes (Wilmut, W-2DPH 1.40 ml, Nirco, S.L., Barcelona, Spain). The first tube was standardized to a minimum volume of 150 µl and was identified with a V in the upper part of the tube, where there was no plasma so as not to interfere with color visualization. All tubes with plasma content were reserved by storing them at -80 ° C. Subsequently, the labeled samples were thawed by slow thawing with an orbital shaker. Immediately prior to spectrophotometric analysis, the plasma-containing tubes were vortexed.

Each observer was asked to visually inspect the color tone of the plasma in a tube of each isolated blood sample, to evaluate and record the stage of the degraded scale with which the tone of the plasma contained in the tube coincided, according to their criteria. . Their agreement and discordance were verified according to the stage corresponding to the absorbance measured for the plasma of that sample.

Traceability between numbers and letters was developed. The first observation of the plasma tubes corresponded to numbers and the second to letters. The plasma tubes were arranged from left to right in ascending numerical and alphabetical order, respectively. The traceability record was made by a person responsible for coordinating the field work, who had single access. With the identification of the second alphabetical sequence, the observers were blinded. The absorbance measurement was blind and was performed in a different laboratory than the laboratory where the visual evaluation was performed.

A minimum number of 6 and a maximum of 16 samples per batch was considered adequate and the procedure was performed twice per sequence in each batch of 1.4 ml tubes:

- test: the centrifuged tube was collected and 150 µl of the total plasma was transferred to a tube of 1.4 ml. The sample contained in the tube was visually evaluated and subsequently frozen;

- re-test: 24-48 hours after freezing, the tubes with plasma content used in the test were thawed and those samples were visually evaluated again; subsequently, they were analyzed by spectrophotometer, as indicated in example 3, subsequently freezing the tube.

Artificial light was used to perform the visual evaluation.

The person responsible for coordination monitored the non-influence between observers, preventing them from coming together in the same laboratory. Each evaluation record was stored, immediately after it was made, in an envelope with limited access only to the coordinator and kept under lock and key.

4.1. Sample size.

A one-sided sample calculation of the observed proportion with respect to a reference was carried out, with an alpha error of 5% and a power of 80%, one-sided test. The reference proportion was 46.6% based on a study in which, of 15 samples identified with the presence of hemolysis through spectrophotometry, 8 of them were not visually detected and 7 could be detected through the human eye (Shash , JS et al. PLoS ONE 2016; 11 (4) 1-12). The objective was to achieve a concordance of at least 85%, justified by being a value of certainty considered sufficiently satisfactory; which was considered optimal for detecting the intensity of hemolysis at each stage.

H0: Hemolysis intensity concordance across visual stages is not optimal <0.70.

H1: Hemolysis intensity concordance across visual stages is optimal ³ 0.85; at least 0.85.

Therefore, the minimum difference to detect was obtained through the objective of 85% - 46.6%, that is, 38.4%. Finally, it was concluded that at least 9 samples were needed or, what is the same, 10 blood samples per stage, with an expected loss proportion of 10%, mainly due to samples with insufficient blood content for inspection. Since the scale consisted of 7 stages, in total, 63 isolated blood samples were needed per scale, or 70 blood samples per scale, including losses.

4.2. Statistical analysis.

The parameters of reliability, validity and feasibility were evaluated.

The Spearman correlation coefficient was used to observe the trend and strength of the relationships between variables; Cohen's Kappa index to know the concordances and disagreements of the stages; the Kaiser-Meyer-OIkin (KMO) test and Bartlett's sphericity test (chi-square model); the intraclass correlation coefficient for the reliability of spectrophotometry on all repeated absorbance measurements.

A nonparametric distribution was detected through the Kolmogorov-Smirnov normality test (p-value <0.05). Therefore, the Kruskal-Wallis statistical test was used to determine the heteroscedasticity between the stages of the scale; and a descriptive statistic to evaluate the feasibility.

The analysis was carried out by compliance with the protocol, the samples necessary to complete the sample size of each stage. That is, those plasma samples that the observers considered were not optimal for visual detection by potential lipemia were excluded.

The SPSS version 23 statistical program and the Microsoft Excel XLSTATA program were used to address the validity of predictive criteria.

Ethical and quality aspects

The subjects signed the informed consent corresponding to a clinical trial in which they participated. From the remaining blood to that required by the test, part was used for the elaboration and testing of the scale. However, no traceability was collected in the identification of the blood samples for validation, they were identified in chronological order of obtaining. In addition, no additional blood sample was drawn, plasma was obtained from the amount of blood remaining in the sample justified by the clinical trial for which the subjects provided their signed consent, and was destroyed as appropriate, therefore not considered additional specific consent is required.

The methodology used to carry out the validation of the scale is supported by the STARD diagnostic precision study guide published in 2015 (Bossuyt, P.M. et al. BMJ 2015; 351: h5527).

4.3. Theory / calculation.

At 414 nanometers (nm), where the maximum intensity of free hemoglobin is generally found in plasma, it is a specific measurement point where the absorbance curves of bilirubin and lipemia overlap; between 400 nm and 540 nm for the detection of jaundice, the maximum expression is evident at 460 nm, and between 300 nm and 750 nm for lipemia, without maximum peak but with optimal absorbance at 300 nm. In addition, there are currently some guidelines for the evaluation of cardiovascular risk, performing certain blood tests under conditions not necessarily fasting (such as the measurement of triglycerides and glycosylated hemoglobin), so it is considered that the correction adjustment method could be justified of lipemia that Appierto, V. et al. propose in Bioanalysis. 2014; 6: 1215-1226: (A414-

A385) +0, 16xA385.

In each measurement, the entire range from 0 nm to 750 nm was obtained from each sample. For the elaboration and validation of the scale, three measurements were made corresponding to each plasma sample, the absorbance at 750 nm was calculated to carry out a basal correction in each sample. Also, the absorbance at 414 nm and at 385 nm was calculated. Subsequently, the calculation was performed according to the formula of Appierto, V et al. (2014).

In cases where the 3 absorbance measurements were very different, up to a maximum of 6 additional measurements were made. That is, a minimum of 3 and a maximum of 9 measurements were made on the same plasma sample. Of all the measurements, the arithmetic mean was performed with the 3 most similar measurements, eliminating the most extreme values.

As a result, three corrected absorbance values were obtained. The arithmetic mean of these three corrected measurements was calculated which resulted in a final single value corresponding to one sample.

Number of samples

80 blood samples were included, in total, from 64 subjects. Each stage was validated by a certain number of samples: stage 1 validated by 14 samples, stage 2 validated by 16 samples, stage 3 validated by 10 samples, stage 4 validated by 10 samples, stage 5 validated by 9 samples, stage 6 by 9 samples and stage 7 by 12 samples.

Interobserver reliability

A Cohen Kappa index of 0.491 was obtained in the test and 0.270 in the retest. It is considered a moderate and acceptable agreement, respectively, according to Landis and Koch (Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. Fam Med. 2005 May; 37 (5): 360-3).

Intra-observer reliability

On the other hand, a Cohen's Kappa index of 0.474 was obtained in the test and retest of the first observer. Cohen's Kappa index of 0.489 in the test and retest of the second observer. It is considered a moderate agreement in both observers according to Landis and Koch. Instrument reliability

The intraclass correlation coefficient (ICC) of the repeated absorbance measurements of each sample was determined to represent the stability of the reference instrument measurements. In the re-test, an ICC of 0.995 was obtained in a 95% confidence interval (CI) (0.992 - 0.996).

Internal consistency reliability

Heteroscedasticity between the stages that make up the scale is explained, according to the magnitude of hemolysis that each stage represents. It is observed in table 3 and in figure 3, through the Kruskal-Wallis test (p = 0.000); Each stage contains a different data variance, in such a way that the lower stages are made up of smaller variances and, as the magnitude of hemolysis increases, the variance of the data that make up the successive stages increases according to the close relevance between stages 1 and 2, which dissipates as the sample becomes more hemolyzed.

Table 3. Kruskal-Wallis test for the seven stages of the hemolysis scale.

In combination, Table 4 represents a significant increase in absorbance across the stages; through a Spearman correlation:

a The null hypothesis is not assumed.

b Use of the asymptotic standard error that the null hypothesis assumes.

c Based on normal approximation.

Table 4. Relationship between arithmetic mean of absorbances and stages. Spearman correlation. In the same way, the standard deviations of the plasma samples used for the validation of each stage are shown, which were the following: Stage 1: 0.018; stage 2: 0.009; stage 3: 0.013; stage 4: 0.015; stage 5: 0.055; stage 6: 0.036; and stage 7: 0.277.

As a consequence of the pairs of stages of possible combinations for their comparisons, an alpha adjustment was carried out through the Bonferroni test. As there were 21 combinations in total, the fixed alpha was raised from 0.05 to 21. A significant alpha <0.002 and a confidence interval of 99.8% were obtained.

Appearance validity

The magnitude and clinical relevance of hemolysis represented is supported by underlying theory in a panel of experts (Appierto, V. et al. Bioanalysis. 2014; 6: 1215-1226; Saldaña, IM An Fac Med 2015; 76 (4): 377- 384; Plumhoff EA, et al. Mayo Medical Laboratories. Communiqué. 2008; 33 (12): 1-8).

Content validity

The adequacy of the absorbances of the plasma samples obtained in the experimental field work was evaluated to reflect the full spectrum of the magnitude of hemolysis absorbance, as shown in Table 5, Kaiser-Meyer-OIkin (KM O) ³ 0.70. In the same way, in Table 5, it was evaluated that the variables of arithmetic mean of absorbance and the variable of stages keep a significant correlation between both to represent the dimension of hemolysis (p-value = 0.000). Therefore, the dimension was correctly represented and there were no other dimensions, unforeseen, responsible for the magnitude of hemolysis in the plasma samples from the field work.

Table 5. KMO and Bartlett tests of the stages according to arithmetic means. Content validity.

Criterion validity. Global concurrent validity

The concordances of the stages of the scale were evaluated globally. In other words, concordances of stages assigned by observers in the test and retest were evaluated in reference to the stage assigned by the arithmetic mean of the absorbance measurements of each plasma sample. Table 6 shows a global concordance ³ 70% in the re-test.

Table 6. Global agreement and disagreement of the stages according to the arithmetic mean of absorbance as a function of the test and re-test moments.

Absolute intra-stage agreement of the scale in the retest: stage 1 of 100%, stage 2 of 100%, stage 3 of 90%, stage 4 of 60%, stage 5 of 33.3%, stage 6 of 11, 1 % and stage 7 of 63.6%. In the discrepancies found, a Spearman Rho was obtained between the visual stage of the observers and the one prepared by absorbances of 90%; the correlation was significant at the 0.01 level (one-sided).

Predictive validity.

It is observed in table 6 that the test does not reach a satisfactory parameter, evaluating the plasma samples in the 48 h before being analyzed. However, in the re-test, in which a satisfactory parameter is found, it is also predictive, since it is evaluated before analyzing absorbances, but in a short period of time. Therefore, table 7 shows the global criterion measurements of the scale in the re-test, that is, from stages 2 to 7 with respect to stage 1. The stated concordances refer to correct absolute classifications with implicit clinical relevance . The prevalence of stage 1 was calculated to be 0, 188.

Stage 2: Concordances of 0.886 in a 95% confidence interval (CI) (0.780-0.780), discrepancies of 0.1 14 95% CI (0.009-0.220), sensitivity of 0.80 in a 95% CI (0.577 -0.923), specificity of 1.00 in a 95% CI (0.757-1.00), false positive rate of 0.00, false negative rate of 0.200 in a 95% CI (0.040-0.360). Prevalence of 0.571 in a 95% CI (0.407-0.735), Positive Predictive Value (PPV) of 1.00 in a 95% CI (1.00-1.00), Negative Predictive Value (NPV) of 0.952 in a CI 95% (0.857-1.00). Negative likelihood ratio (LR-) of 0.20 in a 95% CI (0.083-0.481). Relative risk (RR) of 4.750 in a 95% CI (2.1 12-10.681).

Stage 3: Concordances of 0.960 in 95% CI (0.883-1.00), discrepancies of 0.040 in 95% CI (0.000-0, 117), sensitivity of 0.90 in 95% CI (0.571-1, 000), specificity of 1.00 in a 95% CI (0.757-1.00), false positive rate of 0.00, false negative rate of 0.100 in a 95% CI (0.000-0.257). Prevalence of 0.400 in a 95% CI (0.208-0.592), PPV of 1.00 in a 95% CI (1.00-1.00), NPV of 0.986 in a 95% CI (0.928-1.00). LR- of 0.100 in a 95% CI (0.016-0.642), RR of 16.000 in a 95% CI (3.479-73.583).

Stage 4: Concordances of 0.600 in 95% CI (0.438-0.762), discrepancies of 0.400 in 95% CI (0.238-0.562), sensitivity of 0.300 in a 95% CI (0, 145-0.522), specificity of 1,000 in a 95% CI (0.757-1.00), a false positive rate of 0.000, a false negative rate of 0.700 in a 95% CI (0.517-0.883). Prevalence of 0.571 in a 95% CI (0.407-0.735), PPV of 1.00 in a 95% CI (1,000-1,000), NPV of 0.909 in a 95% CI (0.804-1, 00). LR- of 0.700 in a 95% CI (0.525-0.933), RR of 2.071 in a 95% CI (1.435-2.990).

Stage 5: Concordances of 0.947 in 95% CI (0.847-1, 000), disagreements of 0.053 with 95% CI (0.000-0, 153), sensitivity of 0.750 in a 95% CI (0.290-0.960), specificity of 1, 000 at a 95% CI (0.757-1.00), false positive rate of 0.000, false negative rate of 0.250 at a 95% CI (0.000-0.550). Prevalence of 0.211 in a 95% CI (0.027-0.394), PPV of 1,000 in a 95% CI (1.00-1.00), NPV of 0.969 in a 95% CI (0.884-1, 000), LR- of 0.250 in a 95% CI (0.046-1, 365), RR of 16,000 in a 95% CI (3.479-73.583).

Stage 6: Concordances of 0.842 in 95% CI (0.678-1.000), discrepancies of 0.158 in 95% CI (0.000-0.322), sensitivity of 0.250 in a 95% CI (0.040-0.710), specificity of 1.000 in a 95% CI (0.757-1.00), a false positive rate of 0.000, a false negative rate of 0.750 in a 95% CI (0.450-1,000). Prevalence of 0.211 in a 95% CI (0.027-0.394), PPV of 1,000 in a 95% CI (1,000-1,000), NPV of 0.913 in a 95% CI

(0.782-1, 000), LR- of 0.250 in a 95% CI (0.046-1, 365), RR of 6,000 in a 95% CI

(2,336-15,411).

Stage 7: Concordances of 1,000 in 95% CI (1,000-1,000), discrepancies of 0,000 in 95% CI (0,000-0,000), sensitivity of 1,000 in 95% CI (0.590-1,000) ), specificity of 1,000 at 95% CI (0.757-1, 000), false positive rate of 0.000, false negative rate of 0.000 at 95% CI (0.000-0.000). Prevalence of 0.318 in a 95% CI (0.124-0.513), PPV of 1,000 in a 95% CI (1,000-1,000), NPV of 1,000 in a 95% CI

(1,000-1,000), LR- of 0.000 at a 95% CI (0.000-0.000). It was not possible to calculate

RR.

Construct validity

A proportion of 0.713 in a 95% confidence interval (0.611-0.814) of global scale agreement is considered to be a sufficiently satisfactory parameter despite not rejecting the null hypothesis. In addition, the PPVs of the stages of the scale exceed 85%; being a more practical measurement for the objective designed in practical clinical application. Feasibility

The observers dedicated a visual inspection time in each sample represented by arithmetic mean (variance) of 0.200 (± 0.003) minutes, minimum of 0.12 minutes and maximum 0.33 minutes in the re-test (number of samples = 80) .

The observers required prior training in the use of the 10-minute scale.

4.4. Interpretation of results

The global reliability between the observers is considered acceptable, and the intra-observer reliability moderate, according to Landis and Koch, in the re-test of the scale, which suggests a slight variation in judgment between the observers, which may be due to a lack of unification. criteria.

Repeated absorbance measurements in samples represent stability based on the intraclass correlation index represented, approximately 99%, both of the plasma content to be measured and of the spectrophotometer measuring instrument.

A significant heteroscedasticity is evidenced between the stages through Kruskal-Wallis (p = 0.000). In such a way that the higher the stage number, the greater the variance of the absorbances of the included plasma samples. It is consistent with the clinical relevance that most of the parameters measured in plasma begin to trigger cumulative alterations in stage 2 and successive, with a very narrow variance between stages 1 and 2, so the scale is consistent with the representation of a sample of plasma in a specific stage and not in several; therefore it is discriminatory. Additionally, a strong relationship is observed because as the value of the stage from 1 to 7 increases, the value of the absorbances increases. Therefore, the scale is considered to be coherent with the theoretical measurement construct, its magnitudes and clinical relevance, since heteroscedasticity and the strong correlation between absorbances and an increase in the number of stages are expected.

The plasma samples included in the fieldwork of the scale cover a sufficient spectrum to test this scale (KMO = 0.71). That is, it is based on an accumulation of possible cases that represent the majority of the cases that can be expected when applied in another environment; in which case, to be able to identify each case and expose it through the classification from 1 to 7. In the same way, the unidimensionality of the scale elaborated by the stage variables and arithmetic mean of absorbances, which reveals the correct approach to the hemolysis construct and its magnitudes in a single dimension.

With regard to criterion validity, the overall proportion of concordances is considered satisfactory. These concordances are distributed non-proportionally among their stages, but, with the exception of stage 4, all the others reach a satisfactory value (greater than or equal to 70%). Stage 4, despite not achieving a satisfactory result, not only results in a close value, but the value of 0.70 is among the possible values of concordances collected in the interval of stage 4 in the sample studied with a 95% probability of success. Furthermore, stage 4 maintains a strong relationship between visual assessment and absorbance analysis, suggesting a robust orientation of potentially altered parameters upon reaching the intensity of hemolysis corresponding to stage 4. All stages are useful for the detection of absence of hemolysis with clinical impact; 100% specificity and 0% false positives. Sensitivity ³ 70% except stages 4 and 6. The false negative rate is small, except for stages 4 and 6. All stages predict the corresponding amount of hemolysis, adequately (PPV ³ 70%). Despite the low prediction of the absence of hemolysis in plasma samples globally (NPV £ 70%), probably influenced by the low prevalence, in stages 2 to 7 included it is satisfactory (NPV ³ 70%). The likelihood ratio that a hemolyzed sample is not identified as such and its magnitude is low, except in stage 4, which has not found a discriminatory characteristic. Furthermore, as the stages increase, the probability of screening in the identification of hemolyzed versus non-hemolyzed samples is greater, although the increase is not linear.

Concordances of 71.3% are obtained in the construct validity in a 95% confidence interval of (61.1% -81.4%), the 85% prefixed between its values from the point of view of the global scale. In each of the stages, all reach a construct validity, except stage 4; likely responsible for not reaching global construct validity. For this reason, it is not possible to reject the null hypothesis, it is maintained.

All correct classifications contain at least 70% of their 95% confidence interval values; which corresponds to a classification considered theoretically as adequate.

The scale reflects a criterion validity, not a construct validity, for the sample size approached; it could be due to insufficient capacity in field work. However, it is closely related to the upper limit of the confidence interval of such calculated value.

Limitations

The covariance between stages is not calculated, since all stages are delimited by the immediately consecutive absorbance number; so it does not apply.

It is not possible to know the positive likelihood ratio due to the presence of complete specificity in the field work.

It is not possible to know the construct validity through a hypothesis test because it is one-sample and, therefore, the p-value does not apply.

It is not possible to calculate a divergent validity or reliability of non-bioequivalence between the scales, since it does not apply.

Nor can the discriminative validity be known in lipemia and bilirubinemia, because possible interferences have been eliminated through analytical tests for a bilirubin control in the normal range and extraction of samples in the fasting state, preferably, in addition to applying the correction factor of lipemia.

conclusion

In the detection of visual hemolysis, with bilirubin in the normal range controlled and corrected in lipemia, the degraded scale of hemolysis intensity is a useful tool that reflects the reality of the underlying theoretical construct with a sufficiently satisfactory impact of clinical benefits.

Claims

1. Procedure for the visual detection of the degree of suitability of an isolated blood sample for its validation in clinical analysis based on the degree of hemolysis of the sample, which includes:

- obtaining plasma from the sample,

- the comparison of the color of the plasma of the sample with a scale degrades color that includes the whole range of color intensities from 55 ° to 00 ° and from 00 ° to 351 °, with 00 ° being the color red in the chromatic circle of a double cone and in which said range of intensities is divided into 7 intervals of equal size that correspond to the following stages: stage 1: 55 ° -44 °; stage 2: 43 ° -38 °; stage 3: 37 ° -26 °; stage 4: 25 ° -16 °; stage 5: 15 ° -07 °; stage 6: 06 ° -01 °; stage 7: 00 ° -351 °, whose harmonization in saturation and brightness, respectively, are: stage 1: 35% - 44%, 99% - 89%; stage 2: 47% -50%, 90% -90%; stage 3: 50% -60%, 93% - 89%; stage 4: 68% -80%, 95% - 80%; stage 5: 71% -81%, 91% -78%; stage 6: 68% -83%, 80% -

70%; stage 7: 82% -88%, 70% - 55%,

- the assignment of the sample to one of the 7 stages of the degraded scale.

2. Method according to claim 1, in which the color gradient scale occupies a surface 210 mm high by 297 mm long, each stage corresponding to a size of 105 mm high by 41 mm wide, the set of The colors of the scale are 105 mm high by 287 long, with a margin of 52.5 mm at the top as well as at the bottom and 5 mm at the right and left, in which the color is matte white.

3. Method according to any of the preceding claims, wherein the device where the color degraded scale is included is made of a material comprised in the following group: paper, cardboard, pen-board, plastic and / or metal.

4. Graduated color scale for the visual detection of the degree of suitability of an isolated blood sample for its validation in clinical analyzes based on the degree of hemolysis of the sample, according to the procedure defined in claims 1-3, which includes all the range of color intensities from 55 ° to 00 ° and from 00 ° to 351 °, where 00 ° is the color red on the color wheel and in which said range of intensities is divided into 7 equal intervals size that correspond to the following stages: stage 1: 55 ° -44 °; stage 2: 43 ° -38 °; stage 3: 37 ° -26 °; stage 4: 25 ° -16 °; stage 5: 15 ° -07 °; stage 6: 06 ° -01 °; stage 7: 00 ° -351 °. Whose harmonization in saturation, and brightness respectively is: stage 1: 35% - 44%, 99% - 89%; stage 2: 47% -50%, 90% -90%; stage 3: 50% -60%, 93% - 89%; stage 4: 68% -80%, 95% - 80%; stage 5: 71% -81%, 91% -78%; stage 6: 68% -83%, 80% - 70%; stage 7: 82% -

88%, 70% - 55%.

5. Device for the visual detection of the degree of fitness of an isolated blood sample for its validation in clinical analysis based on the degree of hemolysis of the sample that includes the color scale defined in claim 4, wherein the degraded scale occupies an area of 210 mm high by 297 mm long, each stadium corresponding to a size of 105 mm high by 41 mm wide, the set of colors in the scale occupies 105 mm high by 287 long, with a margin of 52.5 mm in the upper part as in the lower part and 5 mm in the right part as the left part in which the color is matt white.

Device according to claim 5 made of a material selected from the group: paper, cardboard, foam board, plastic and / or metal.