CC BY
42
UDC 616.133.33:616.152.21-07
THE ALGORITHM OF LABORATORY TESTS TO PREDICT CEREBRAL LESIONS BY PERINATAL HYPOXIA
1 Altai State Medical University, Barnaul
2Federal center of traumatology, orthopedics and endoprosthesis, Barnaul
Yu.V. Shabalina1, S.A. Yelchaninova1, A.V. Popovtseva12, Yu.V. Degtyareva1, Yu.V. Korenovsky1
In a randomized prospective clinical study involving 150 women and newborns there are evaluated the operational characteristics of the umbilical cord blood tests in relation to the differential diagnosis of perinatal hypoxic cerebral lesions in newborns. The method of multidimensional scaling revealed that the algorithm based on the combination of the levels of matrix metalloproteinase-9 (MMP-9), thiobarbiturate-reactive substances (TBARS), erythropoietin and lactate in the first hours after birth allows to predict perinatal hypoxic cerebral lesions with a sensitivity of 85% and a specificity of 90%. Key words: perinatal hypoxia, laboratory biomarkers.
Perinatal hypoxia (PX) is a pathological state of the perinatal period developing due to the lack of oxygen delivery to the organism of the fetus and/or its uptake during pregnancy (chronic in-trauterine hypoxia), delivery (acute fetal hypoxia, asphyxia), early neonatal period [1, 2]. PX complication is the multiple organ damage in the fetus and newborn which manifests itself through various nosological forms [3]. Herewith, 15-20% of all PX complications both in mature and premature newborns constitute hypoxic-ischemic encepha-lopathy and intraventricular hemorrhage. These central nervous system (CNS) damages by PX are revealed according to neurological symptoms and the data of ultrasound investigation, as a rule, not earlier than on the 3rd-4'h day of age [4-6].
Diverseness and remoteness of the manifestation of clinical and neurovisualization symptoms of PX complications from the moment of birth considerably complicate timely diagnosis and, consequently, the effectiveness of prevention and correction of target organ lesions, including CNS damages [7, 8]. This stimulates the search of biochemical markers for the prognosis and early diagnosis of perinatal hypoxic damages [9, 10].
It is known, that the key stages of PX pathogen-esis are such metabolism disorders induced by hy-poxia- reperfusion and interrelated systemic disorders as the increase of lactic-acid fermentation, free-radical oxidation, increase of matrix metallo-proteinases activity [11-14]. As the compensatory reaction by PX is considered the growth of erythro-poietin production [15, 16].
In this regard, the research objective was
to elaborate the algorithm of prognosis of CNS damages in newborns by PX on the basis of complex of pathogenetic laboratory indexes of systemic metabolic disorders.
Materials and methods
The randomized prospective clinical study included 150 pregnant women and their newborns
separated into two groups depending on the presence/lack of symptoms of perinatal hypoxic CNS lesions in newborns at the early neonatal period.
The exclusion criteria were: immunoincompat-ible and multifetal pregnancy; delivery time less than 28 weeks; presence of reproductive disorders in the medical history of a woman, extragenital diseases at the stage of subcompensation and decompensation, preeclampsia; drug and/or toxic (including alcohol) dependence, diabetes mellitus; presence of congenital abnormalities and genetic metabolic disorders in the newborns.
The formed groups of mothers of the newborns did not differ in the age, social status, obstetric and gynecological anamnesis, presence of chronic extragenital diseases. The clinical characteristics of the newborns is presented in Table 1.
PX was diagnosed on the basis of presence of not less than two following symptoms: intranatal distress with fetal bradycardia, 100 and more beat per minute according to cardiotocography; 6 and less points according to the Apgar scale on the 5th minute; requirement of resuscitation for more than 1 minute [4]. CNS lesions were evaluated on the basis of symptoms of hypoxic-ischemic changes in the brain according to the data of ultrasound investigation on the 3rd-4'h day of birth. Newborns with PX were transferred to the department of intensive therapy, where there were conducted essential resuscitation procedures including artificial lung ventilation (ALV).
In the newborns there was performed the standars laboratory examination (general blood analysis, differential leucocyte count, the level of bilirubin, urea, creatinine, crude protein, aminotransferase, alkaline phosphatase, glucose, lactate, potassium, sodium), screening examination for phenylke-tonuria, cystic fibrosis, hypothyrosis, adrenogen-ital syndrome and galactosemia. For the specific tests immediately after the birth there was taken venous blood from the umbilical artery into vacutainers with lithium heparinate (14 UN per
Table 1
Clinical characteristics of the newborns
Index Newborns without perinatal hypoxic CNS damage n = 62 Newborns with perinatal hypoxic CNS damage n = 40 Significance of differences, P
Delivery term 39,3 ± 4,95 34,8 ± 3,1 < 0,001
Weight (kg) 3379,0 ± 461,38 2415,5 ± 892,97 <0,001
Apgar 1st min 7,2 ± 0,82 6,000 ± 0,96 <0,001
Apgar 5th min 8,1 ± 0,68 7,1 ± 0,55 <0,001
ALV, n (%) ALV - mask, n (%) 5 (8,1) 0 (0) 24 (60) 3 (7,5) < 0,001 < 0,001
Respiratory distress-syndrome, n (%) 4 (6,4) 30 (75) < 0,001
Caesarean operation, n (%) Amniotomy, n ( %) 28 (90,3) 2 (6,5) 23 (88,5) 2 (7,7) 0,974 0,974
Syndrome of CNS damage: Distress, n (%) Activation, n (%) 3 (4,8) 1 (1,6) 34 (85) 2 (5) < 0,001 < 0,001
Type of CNS damage: Hypoxice, n (%) 0 (0) 8 (20) < 0,001
Ischemic-hypoxic, n (%) 4 (6,4) 28 (70) < 0,001
Note: quantitative variables are presented in the form of M (arithmetic mean) ± SD (standard deviation).
1 ml blood), and also into vacutainers with sodium fluoride and K2-EDTA (3,3 and 1,2 mg per 1 ml blood respectively).
In the blood serum by means of the enzyme immunoassay there was measured the concentration of matrix metalloproteinases (MMP-1, MMP-2, MMP-9), tissue inhibitor of matrix metallopro-teinases (TIMP-1) using test systems produced by "RayBiotech" (USA); erythropoietin, S100B protein using test systems produced by "Biomer-ica" (USA) and "CanAG" (Sweden) respectively; by means of absorption photometry there was measured the concentration of lactate with reagents "Biocon" (FRG) and products of lipid peroxidation of thiobarbiturate-reactive products (TBRP) with reagents "ZeptoMetrixCorporation" (USA).
The study is approved by the Ethics Committee of the FSBEI HE Altai State Medical University of the Ministry of Health of the Russian Federation, performed on the clinical base of RSBHI Altai Krai Perinatal (clinical) center (Barnaul) in the period from 2008 to 2012.
By the statistical analysis of the study results by means of SigmaPlot 11.0 program there was evaluated the normality of distribution of variables according to Shapiro-Wilk test, intergroup differences of quantitative variables according to Mann-Whitney U test, categorical variables according to chi-square. Mathematical modelling by means of multidimensional scaling, visualization of the diagnosis algorithm and associated calculations were performed by means of JMP 7.0 program (SAS Institute, USA) in the mode Partition, ROC-analysis (receiver operating characteristic analysis) - in MedCalc 9.1.0.1 program. The critical level for all used criteria was considered p < 0,05.
Results and discussion
According to the analysis, the ROC-curve distribution of the values of all studied parameters, except for TIMP-1 concentration, and also the calculated rates (relation of the level of stated matrix metalloproteinases to the level of TIMP-1) differed significantly from the distribution with the area 0,5 (Table 2). This indicates the possibility of using these indexes as biomarkers to prognosticate CNS disorders by PX. The biggest area under ROC-curve is determined for erythropoietin, TBRP, lac-tate, MMP-2 and relation MMP-2/TIMP-1.
Table 3 presents operational characteristics of the studied parameters. According to the data, the combination of high sensitivity and high specificity in terms of CNS lesions in the newborns with PX rational for clinical practice was not revealed for any of the examined indexes. Consequently, it is logical to search for the combination of tests with high sensitivity and tests with high specificity.
In order to solve the stated problem, there was carried mathematical modelling of the laboratory algorithm of prognosis of cerebral disorders in the newborns by perinatal hypoxia. In terms of modelling there was stated that, according to the rate of true-positive and true-negative results, the best result of differentiating the newborns according to the studied pathology is achieved by using the levels of TBRP, lactate, MMP-9 and erythropoi-etin. The quantitative evaluation of the combination of these tests showed, that sequential analysis of the levels of MMP-9, TBRP, lactate and erythro-poietin (Fig.) allows to reveal 85,0% of newborns with perinatal hypoxic CNS lesion (sensitivity 85%) and 90,0% of newborns without such pathology (specificity 90%) already in the first hours of birth.
Area under ROC-curve for tests in the venous blood plasm from the umbilical artery of the newborns in relation to the prognosis of cerebral damages by perinatal hypoxia lauie z.
Test n Area under ROC-curve (CI95%) P
Erythropoietin 56/85 0,807 (0,732—0,869) <0,0001
TBRP 57/87 0,807 (0,733—0,868) <0,0001
Lactate 55/85 0,782 (0,705—0,848) <0,0001
MMP-2 55/84 0,699 (0,616—0,774) <0,0001
MMP-2/TIMP-1 51/79 0,679 (0,591—0,758) 0,0007
MMP-1 58/85 0,672 (0,589—0,748) ab 0,0004
MMP-1/TIMP-1 53/78 0,667 (0,580—0,747) abc 0,0005
MMP-9 54/84 0,646 (0,560—0,726) abc 0,0099
S100B protein 59/84 0,610 (0,525—0,690) abc 0,0304
MMP-9/TIMP-1 50/81 0,610 (0,521—0,694) abc 0,0368
TIMP-1 54/85 0,527 (0,440—0,612) abc 0,5903
Note: CI95% — 95% confidence interval; n — number of newborns with cerebral lesion by perinatal hypoxia/number of newborns without lesions; p — significance of difference of the area under ROC-curve in comparison with 0,5; abc — p<0,05 for the area under ROC-curve in relation to the same value for TBRP, erythropoietin and lactate respectively.
Table 3 Operational characteristics of tests in the venous blood plasm from the umbilical artery of the newborns in relation to the prognosis of cerebral damages by perinatal hypoxia
Test Border Sp (CI95%) S (CI95%) +LR/-LR +PV/-PV
Erythropoietin, mUn/ml >56,2 57,1 (43,2 — 70,3) 91,8 (83,8—96,6) 6,94/0,47 82,1/76,5
TBRP, mmol/l >31,9 75,4 (62,2—85,9) 79,3 (69,3—87,3) 3,65/0,31 70,5/83,1
Lactate, mmol/l >8,5 63,6 (49,6 — 76,2) 82,4 (72,6—89,8) 3,61/0,44 70,0/77,8
MMP-2, ng/ml <49,2 47,3 (33,7—61,2) 95,2 (88,3—98,7) 9,93/0,55 86,7/73,4
MMP-2/ TIMP-1 <1,98 51,0 (36,6—65,2) 86,1 (76,5—92,8) 3,66/0,57 70,3/73,1
MMP-1, ng/ml >305,7 62,1 (48,4— 74,5) 71,8 (61,0—81,0) 2,20/0,53 60,0/73,5
MMP-1/ TIMP-1 <0,014 71,7 (57,7—83,2) 57,7 (46,0—68,8) 1,69/0,49 53,5/75,0
MMP-9, ng/ml >107,1 50,0 (36,1—63,9) 95,2 (88,3—98,7) 10,50/0,53 87,1/74,8
Protein, S100B, ng/l >0,99 45,8 (32,7—59,2) 79,8 (69,6—87,7) 2,26/0,68 61,4/67,7
MMP-9/ TIMP-1 >4,16 44,0 (30,0—58,7) 82,7 (72,7—90,2) 2,55/0,68 61,1/70,5
TIMP-1, ng/ml >21,9 81,5 (68,6—90,7) 34,1 (24,2 —45,2) 1,24/0,54 44,0/74,4
Note: Sp — specificity; S — sensitivity; +LR accuracy of determination of positive results;
— positive relation -PV — accuracy of
of chances; -LR — negative relation of chances; +PV — determination of negative results.
It should be noted, that by modelling the algorithm on the basis of maximum coefficient of separation into the groups "disease - no disease" somehow change the borders of tests included into the algorithm in relation to the borders stated by ROC-analysis (Fig. 1 and Table 3). This, in its turn, causes the change of sensitivity and specificity of tests and also the creation of cycles - backup to the tests of the previous level by certain results. At the same time, the cyclic character of the algorithm does not require any additional laboratory investigations, and only complicates a little the interpretation of the results at the postanalytical phase.
It appears that the elaborated algorithm of prognosis of hypoxic cerebral damages in the newborns based on four tests in the plasm of umbilical blood is promising for the validation in prospective studies. In terms of practical implementation of the al-
gorithm rational is the elaboration of express methods of measuring MMP-9 and erythropoietin levels and computer program for result analysis.
References
1. Gomella T.L, Kannigam M.D. Neonatology. Moscow, 2002.
2. Mikhailov A.V., Tunell R. Clinical guidance on the asphyxia of the fetus and newborn. Saint Petersburg, 2011.
3. Gazzolo D., Abella R. et. al. Circulating biochemical markers of brain damage in infants complicated by ischemia reperfusion injury. Cardiovasc. Hematol. Agents Med. Chem. 2009; 7: 108-126.
4. Zakharova L.A. Mutual regulation of the neuroendocrine and immune systems. Ontogenesis. 2010; 6(41): 414-424.
Figure 1.
Algorithm of differential diagnosis of perinatal hypoxic CNS damage in the newborns based
on laboratory tests in the umbilical blood Units of concentration like in Table 3; hatching shows the borders identifying the newborns with perinatal hypoxic CNS damage
10.
11.
Inder T.E., Volpe J.J. Mechanisms of perinatal brain injury. Semin. Neonatol. 2000; 1 (5): 3-16.
Jensen A., Garnier Y., Middelanis J., Berger R. Perinatal brain damage - from pathophysiology to prevention. Eur. J. Obstet. Gynecol. Reprod. Biol. 2003; 110 (Suppl. 1): 70-79. Palchik A.B., Shabalov N.P. Hypoxic-ischemic encephalopathy of newborns. Moscow, 2011. Buonocore G., Perrone S. Biomarkers of hypoxic brain injury in the neonate. Clin. Perinatol. 2004; 1 (31): 107-116. Risso F.M., Sannia A., Gavilanes D.A. et. al. Biomarkers of brain damage in preterm infants. J. Matern. Fetal Neonatal Med. 2012; 4: 101-104.
Tonni G., Leoncini S., Signorini C. Pathology of perinatal brain damage: background and oxidative stress markers, 2014. Korenovsky Yu.V., Yelchaninova S.A. The biochemical markers of hypoxic perinatal affections of central nervous system in newborns. Klinicheskaya Laboratornaya Diagnostika. 2012; 2: 3-7. Dennery P.A. Role of redox in fetal development and neonatal diseases. Antioxid. Redox Signal. 2004; 6: 147-153.
13. Evrard P. Pathophysiology of perinatal brain damage. Dev. Neurosci. 2001; 23: 171174.
14. Kurinczuk J.J. Epidemiology of neonatal encephalopathy and hypoxic-ischaemic encephalopathy. Early Hum. Dev. 2010; 6 (86): 329-338.
15. Maier R.F. Cord blood erythropoietin in relation to different markers of fetal hypoxia. Obstet. Gynecol. 1993; 4 (81): 575-580.
16. Shalak L., Perlman J.M. Hypoxic-ischemic brain injury in the term infant-current concepts. Early Hum. Dev. 2004; 80: 125-141.
Contacts:
Corresponding author - Shabalina Yulia Vadimov-na, Assistant of the Department of general and biological chemistry, clinical laboratory diagnostics of the FSBEI HE Altai State Medical University of the Ministry of Health of the Russian Federation, Barnaul.
656038, Barnaul, Lenina Prospekt, 40. Tel.: (3852) 241392. Email: timidin@mail.com
