|
 
 |
| ORIGINAL ARTICLE |
|
| Year : 2019 | Volume
: 26
| Issue : 4 | Page : 223-229 |
|
The use of uterine artery doppler indices for prediction of pre-eclampsia in Port-Harcourt, Nigeria
Emecheta Gabriel Okwudire1, Omolola Mojisola Atalabi2, Ugonna Micheal Ezenwugo3
1 Department of Radiology, Braithwaite Memorial Specialist Hospital, Port-Harcourt, Nigeria
2 Department of Radio Diagnosis, University College Hospital, Ibadan, Nigeria
3 Federal Medical Centre, Owerri, Nigeria
| Date of Web Publication | 4-Oct-2019 |
Correspondence Address:
Dr. Emecheta Gabriel Okwudire
Department of Radiology, Braithwaite Memorial Specialist Hospital, Port-Harcourt
Nigeria
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/npmj.npmj_54_19
Context: Pre-eclampsia (PrE), a clinical syndrome characterised by elevated blood pressure arising after 20 weeks of gestation, is a leading cause of maternal death worldwide. We evaluated the role of uterine artery Doppler (UtAD) in screening for PrE among unselected, pregnant women. Methodology: This was a prospective cohort study of 170 healthy gravid women between 18 and 26 weeks of gestation recruited consecutively from the Antenatal Clinic of Braithwaite Memorial Specialist Hospital, Port-Harcourt, Nigeria, between July 2016 and June 2017. All had UtAD scans with an abnormal result defined as pulsatility index (PI), resistance index or systolic/diastolic (S/D) ratio >95th centile for gestational age or proto-diastolic notching. Outcome was obtained from antenatal records. Data were analysed using Statistical Package for Social Sciences, version 20 at statistical significance level of P < 0.05. Results: The prevalence of PrE was 7.6%. There was significant association between an abnormal PI (χ2 = 16.29, P = 0.00), S/D ratio (χ2 = 8.55, P = 0.00) and the combined result (χ2 = 11.5, P = 0.007) with subsequent PrE. The highest sensitivity (53.8%) was obtained for the combined result with specificity, negative predictive value (NPV) and positive predictive value of 86.6%, 95.8% and 25%, respectively, area under the curve (AUC) of 0.71 (95% confidence interval [CI]: 0.534–0.871). A normal result had a very high NPV for all indices. The accuracy for the prediction of severe PrE was greater for all indices being highest for the combined result AUC of 0.830 (95% CI: 0.624–1.000; P = 0.01). Conclusion: Abnormal UtAD indices were associated with PrE and may be used in PrE screening.
Keywords: Doppler, pre-eclampsia, pregnancy, receiver operating characteristics curve, uterine artery
How to cite this article:
Okwudire EG, Atalabi OM, Ezenwugo UM. The use of uterine artery doppler indices for prediction of pre-eclampsia in Port-Harcourt, Nigeria. Niger Postgrad Med J 2019;26:223-9 |
How to cite this URL:
Okwudire EG, Atalabi OM, Ezenwugo UM. The use of uterine artery doppler indices for prediction of pre-eclampsia in Port-Harcourt, Nigeria. Niger Postgrad Med J [serial online] 2019 [cited 2023 Mar 30];26:223-9. Available from: https://www.npmj.org/text.asp?2019/26/4/223/268596 |
| Introduction | |  |
Pre-eclampsia (PrE), is defined as new onset of hypertension (>140/90 mmHg) appearing after 20 weeks of gestation and accompanied by proteinuria (>0.3 g/24 h).[1] It is the leading cause of maternal mortality, intrauterine growth retardation (IUGR) and fetal prematurity affecting 5%–10% of pregnancies worldwide.[2] The prevalence of PrE is 5.7% in South-South Nigeria.[3] Various risk factors have been implicated including general factors such as primigravidity, young maternal age, maternal age >35 years, low-socioeconomic class, family history, multiple pregnancy, obesity, chronic hypertension, renal disease, thrombophilia, gestational diabetes and connective tissue diseases.[4]
The pathophysiology of PrE is related to incomplete trophoblastic invasion of the myometrium leading to limited remodeling of spiral arteries during the first and early second trimester.[5] This defective trophoblastic invasion causes placental insufficiency and damage leading to systemic inflammatory response, which in turn leads to vascular endothelial dysfunction and damage.[6] This manifests clinically as hypertension, proteinuria and other systemic responses. It is known that the impaired placental perfusion caused by vascular abnormalities precedes clinical manifestations of PrE.[7]
Although the incidence of PrE in the general obstetric population is relatively low, there is a potential for serious complications such HELLP syndrome (haemolytic anaemia, elevated liver enzymes and low platelet count), eclampsia, coagulopathy, stroke and death in the mother.[8] Newborns affected by IUGR are at increased risk for hypertension, cardiovascular disease and diabetes later in life.[9] From the foregoing, it will be very beneficial to have a screening test that could identify, early in pregnancy, those women who would develop PrE subsequently. This would allow increased surveillance of those at risk, and reduce surveillance for those unlikely to develop the syndrome.[10] Uterine Doppler (UtAD) is a non-invasive tool that can be used to indirectly assess trophoblast development and uteroplacental perfusion and has been considered a useful method for the prediction of PrE.[11]
This study aimed to determine and compare the accuracy of UtAD indices (resistance index [RI], pulsatility index [PI], systolic/diastolic [S/D] ratio and notching [N]) for the prediction of PrE among gravid women in southern Nigeria.
| Methodology | | |
Ethical considerations
Ethical approval for the study was obtained from the Health Research Ethics Committee of Braithwaite Memorial Specialist Hospital, Port-Harcourt (16/02/2015) prior to commencement of the study. Participants were only recruited after the purpose of the study was explained to them, and a written informed consent form was signed. All Doppler scans were done at no cost to the patients.
Study design/location
This was a prospective, cohort study of UtAD US parameters in healthy pregnant women conducted between July 2016 to June 2017 at Braithwaite Memorial Specialist Hospital, Port-Harcourt, the capital city of Rivers State in southern Nigeria.
Study population
Study population included 170 apparently healthy women with singleton pregnancies between 18 and 26 weeks of gestation (last menstrual period [LMP] or early ultrasound determined) recruited by consecutive sampling from the antenatal Clinic (ANC). Women with co-existing medical conditions/history of medical complications (e.g. chronic hypertension and cardiac disease), multiple pregnancy, fetal abnormality and gross obesity precluding transabdominal scanning were excluded.
Sample size determination: For a cohort study, the sample size was calculated using the formula below:[12]
Where;
Zα= the standard normal deviate, taken to be 1.96 for this study and which corresponds to a confidence level of 95%.
Zβ= the power of the test, taken to be 0.8416 for this study and which corresponds to the power of 80%.
P0= the proportion of the participants in the unexposed group who are expected to exhibit the outcome of interest based on a recent study[13] (4%).
P1= the proportion of the participants in the exposed group who are expected to exhibit the outcome of interest (15.2%).
f = the proportion of study participants who are expected to leave the study for reasons other than the outcome under investigation (10%).
Total sample size N = 120 pregnant women.
A minimum sample size of 120 women was needed for the study. However, 170 participants were recruited to increase the accuracy of the study.
Inclusion criteria
- Participant must be normotensive healthy pregnant woman
- Singleton pregnancy with known gestational age, (LMP or early USS determined) 18 weeks and above.
Exclusion criteria
- Co-existing medical conditions (e.g. chronic hypertension, diabetes mellitus and cardiac disease)
- Multiple pregnancy
- Gross obesity precluding transabdominal scanning
- Fetal abnormality
- Refusal to consent.
Data Acquisition: after obtaining signed informed consent, patients were scanned using a GE Logiq P6PRO ultrasound machine equipped with a 2–6 MHz curvilinear probe. A routine obstetric scan was done, then the transducer was placed obliquely in the lower paracervical area and angled medially till the uterine artery was visualised as it crossed the external iliac artery and vein [Figure 1], having originated from the internal iliac artery. The sample volume was set to cover two-third of the diameter of the uterine artery, and angle of insonation was maintained at 0°. Gain, PRF and wall filter were adjusted to obtain clear spectral waveforms with minimal background noise. Pulse-wave Doppler was updated and the image frozen when at least 3 consecutive good-quality waveforms were obtained, and measurements of the peak systolic, end-diastolic and time-averaged mean velocities taken 1 cm distal to the point of apparent cross over with the external iliac vessels.
UtAD PI, RI and S/D ratios (derived automatically by the machine) were recorded on the datasheet. Doppler findings were considered abnormal if any index was > 95th centile or diastolic notching was present. The women were followed till delivery and pregnancy outcomes were obtained from ANC records.
Data analysis
Data were analysed using Data was analysed using statistical package for social sciences (SPSS) version 20, (IBM Corp., Armonk, New York, USA). Comparison of means was done using the Student's t-test while categorical variables were analysed for association using Pearson's Chi-square test for independence. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and likelihood ratios (LR) were calculated from contingency tables. The screening performance of the indices was compared by constructing receiver operating characteristic (ROC) curves and comparing the area under the curves (AUC). P = 5% (P< 0.05) were considered statistically significant.
| Results | | |
Initially, 183 women were recruited for the study, but outcomes were only available for 170 of them, giving response rate of 93%. Their demographic characteristics are shown in [Table 1]. Mean age was 31.6 ± 4.9 years, and about half (48%) of them were nulliparous. Thirteen of these women developed PrE, giving a prevalence rate of 8% as shown in [Figure 2]. The relationship between the Doppler indices and development of PrE was evaluated [Table 2]. There was significant association between the PI test (X2 = 16.29 p = 0.00), S/D test (X2 = 8.55, p = 0.003) and a combined test (X2 = 11.5, p = 0.001) with development of PrE. The relationships between PI, RI as well as notching with development of PrE were not statistically significant. | Table 2: Association between Doppler test results and development of pre-eclampsia
Click here to view |
[Table 3] displays the screening characteristics of uterine Doppler for PrE. The combined test result had the highest sensitivity of 53.8%. Specificity was high for all indices with PI having the highest specificity of 95.5% closely followed by diastolic notching with a specificity of 92.9%. Patients without uterine artery notching were 3 times less likely to develop PrE than counterparts with notching (95% confidence interval [CI]: 1.04–10.76). All the tests had a high NPV for PrE, while PPV was highest (41.7%) for PI >95th centile. | Table 3: Screening characteristics of uterine Doppler indices for pre-eclampsia
Click here to view |
[Figure 3] shows the receptor-operator curves, while [Table 4] shows the corresponding AUC of the various Doppler indices for the prediction of PrE. The combined test had the best predictive performance with an AUC of 0.71 (95% CI: 0.534–0.871) while UtAD RI was least performing with AUC of 0.56 (95% CI: 0.388–0.735) and showed no discriminatory ability (P > 0.05). | Figure 3: Comparison of areas under the curve of Doppler parameters in the prediction of pre-eclampsia
Click here to view |
 | Table 4: Area under the curve for prediction of pre-eclampsia receiver operating characteristic curve
Click here to view |
Five out of the 13 patients had severe PrE [Figure 4]. [Table 5] shows that abnormal PI, RI, S/D and combined test were significantly associated with severe PrE (P< 0.05), with PI > 95th centile having the strongest association χ2 = 14.5, P = 0.00). The screening accuracy of the UtAD indices for severe PrE is shown in [Table 6]. There was an increase in sensitivity for all indices, with the S/D test again having the highest sensitivity (80%), specificity of 87.3%, NPV of 99.3% and PPV of 16.0%. Patients with an abnormal PI result were the most likely to develop severe PrE (Positive LR = 1.33, 95% CI: 0.96–1.83), while patients with a normal S/D ratio were least likely to develop severe PrE (Negative LR = 0.04, 95% CI: 0.01–0.37). | Table 5: Relationship of uterine artery indices with severe pre-eclampsia
Click here to view |
 | Table 6: Screening characteristics of uterine Doppler indices for severe pre-eclampsia
Click here to view |
There was also increase AUC for prediction of severe PrE was compared [Figure 5] and [Table 7] for all indices. A combined test result had the highest AUC of 0.830 (95% CI: 0.624–1.000) while UtA notching had the least AUC of 0.688 (95% CI: 0.383–0.954). However, only the combined test result and abnormal PI result showing significant discriminatory ability (P< 0.05) for severe PrE. | Figure 5: Comparison of areas under the curve of Doppler parameters in the prediction of severe preeclampsia
Click here to view |
 | Table 7: Area under the curve for receiver operating characteristic for prediction of severe pre-eclampsia
Click here to view |
| Discussion | | |
The index study was a prospective cohort study of the utility of UtAD indices in the prediction of development of PrE during the second trimester. The prevalence of PrE was 7.6% and comparable to 8.8% reported by Musa et al.[14] in Jos, but was much higher than 5.6% found by Adokiye et al.[3] in south-south Nigeria. This difference may be due to variation in study design as the study by Adokiye et al.[3] was a retrospective study. It may also be due to a difference in socioeconomic conditions as their study was carried out in a rural setting compared to the index study carried out in a cosmopolitan state capital.
There was a significant association between S/D ratio result and development of PrE (P = 0.003) and is in consonance with the reports by Madazli et al.[15] However, we found higher specificity (88% vs. 78%), with lower sensitivity and PPV (71% vs. 46.2%; 29% vs. 12%) than they reported. These differences may be due to much higher prevalence of PrE (11.5%) in the studied populations as well as a different definition of an abnormal result as S/D >2.6.
Mean RI showed no significant association with the development of PrE. Similar finding was reported by Naategaal[16] in Australia. In contrast, Rampello et al.[17] in Italy reported significant association between RI and PrE. This variation may be due to the fact that this study was in a high-risk population of Caucasian women. Specificity, PPV and NPV were similar, but the sensitivity was lower than reported in others.[18] This may be because these studies used a different cutoff value (RI >90th centile) and were also done in Caucasian populations.
The mean PI >5th centile was strongly associated with the development of PrE (P = 0.00) in the index study. This is similar to the findings by Cooley et al.[19] and Lopez-Mendez et al.[7] For the prediction of PrE, the mean PI alone had a modest sensitivity of 38.5% and PPV of 41.7% respectively, with high specificity 96%. These findings are similar to those reported by Jamal et al.[13] while Noor et al.[20] found a higher sensitivity. Again, the differences in cut-off values and prevalence of PrE may account for these variations.
No significant association between bilateral uterine artery notching and PrE was found. This contrasts with the findings of Asnafi and Hajian[21] in Iran who reported a significant association of PrE with uterine artery notching. This difference may be because their study was done in a high-risk cohort with a higher prevalence of PrE compared to our unselected study population. Uterine artery notching gave sensitivity of 23.1%, specificity of 92.9%, PPV of 21.4% and NPV of 93.6% for PrE. The sensitivity and specificity are similar to findings by Audibert et al.[22] in France and Espinoza et al.[23] in the United States but is lower than those from other studies in low-risk populations.[24],[25] These variations may be due to difference in study design or racial differences as these were done in Caucasian populations.
The accuracy of various indices for prediction of PrE was compared using ROC curves. The combined test result had the highest performance with an AUC of 0.71 (95% CI: 0.534–0.871) This is similar to findings by Lopez-Mendez et al.,[7] who reported that a combined abnormal result (the combination of a proto-diastolic notch and RI + PI >95th centile) gave the highest predictive accuracy. The finding that UtA RI >95th centile performed poorly varied from the findings of Aquilina et al.[26] and Albaiges et al.[27] who reported the best screening method as a combination of mean RI with bilateral notching. This difference may be due to their use of mean RI as cutoff compared to RI >95th centile used in the index study.
Espinoza et al.[23] reported Bilateral UtA notching and/or a mean UtA PI >95th percentile had the highest AUC for the prediction of PrE. This difference may have arisen due to the poor association of uterine artery notching with PrE in our study (AUC = 0.58, P = 0.317) and thus was minimally helpful in combination with other poorly performing indices for identifying patients who are likely to develop PrE.
All Doppler indices except uterine notching were significantly associated with severe PrE. The sensitivity for detection of severe PrE increased from 38.5% to 75% and 23.1% to 60% for PI and RI, respectively and is consistent with findings from previous studies.[15],[27] The accuracy of all the indices were better for prediction of severe PrE as evidenced by the increase in the AUC.
| Conclusion | | |
UtAD showed some usefulness in identifying patients destined to develop PrE and may be incorporated into the routine anomaly scan. Women with a normal second-trimester screening Doppler would not require further assessment while women with abnormal UtAD may benefit from a follow-up Doppler examination, as the persistence of these abnormalities may increase accuracy of the Doppler indices. Further study of the predictive role of uterine Doppler indices among women with high-risk pregnancies in our environment is recommended.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Brown MA, Lindheimer MD, de Swiet M, Van Assche A, Moutquin JM. The classification and diagnosis of the hypertensive disorders of pregnancy: Statement from the International Society for the Study of Hypertension in Pregnancy (ISSHP). Hypertens Pregnancy 2001;20:11-4.  |
| 2. | Backes CH, Markham K, Moorehead P, Cordero L, Nankervis CA, Giannone PJ. Maternal preeclampsia and neonatal outcomes. J Pregnancy 2011;2011:214365. |
| 3. | Adokiye EK, Isreal J, Harry TC, West OL. Factors influencing the prevalence of Preeclampsia – Eclampsia in booked and unbooked patients: 3 years retrospective study in NDUTH, Okolobiri. World J Med Med Sci 2015;3:1-14. |
| 4. | Jido TA, Yakasai IA. Preeclampsia: A review of the evidence. Ann Afr Med 2013;12:75-85. [ PUBMED] [Full text] |
| 5. | Davison JM, Homuth V, Jeyabalan A, Conrad KP, Karumanchi SA, Quaggin S, et al. New aspects in the pathophysiology of preeclampsia. J Am Soc Nephrol 2004;15:2440-8. |
| 6. | Sacks GP, Studena K, Sargent K, Redman CW. Normal pregnancy and preeclampsia both produce inflammatory changes in peripheral blood leukocytes akin to those of sepsis. Am J Obstet Gynecol 1998;179:80-6. |
| 7. | Lopez-Mendez MA, Martinez-Gaytan V, Cortes-Flores R, Ramos-Gonzalez RM, Ochoa-Torres MA, Garza-Veloz I, et al. Doppler ultrasound evaluation in preeclampsia. BMC Res Notes 2013;6:477. |
| 8. | Sibai B, Dekker G, Kupferminc M. Pre-eclampsia. Lancet 2005;365:785-99. |
| 9. | Barker DJ, Osmond C, Golding J, Kuh D, Wadsworth ME. Growth in utero, blood pressure in childhood and adult life, and mortality from cardiovascular disease. BMJ 1989;298:564-7. |
| 10. | Yu J, Shixia CZ, Wu Y, Duan T. Inhibin A, activin A, placental growth factor and uterine artery Doppler pulsatility index in the prediction of pre-eclampsia. Ultrasound Obstet Gynecol 2011;37:528-33. |
| 11. | Cnossen JS, Morris RK, ter Riet G, Mol BW, van der Post JA, Coomarasamy A, et al. Use of uterine artery Doppler ultrasonography to predict pre-eclampsia and intrauterine growth restriction: A systematic review and bivariable meta-analysis. CMAJ 2008;178:701-11. |
| 12. | Bewick V, Cheek L, Ball J. Statistics review 8: Qualitative data – Tests of association. Crit Care 2004;8:46-53. |
| 13. | Jamal A, Abbasalizadeh F, Vafaei H, Marsoosi V, Eslamian L. Multicenter screening for adverse pregnancy outcomes by uterine artery Doppler in the second and third trimester of pregnancy. Med Ultrason 2013;15:95-100. |
| 14. | Musa J, Mohammed C, Ocheke A, Kahansim M, Pam V, Daru P, et al. Incidence and risk factors for pre-eclampsia in Jos Nigeria. Afr Health Sci 2018;18:584-95. |
| 15. | Madazli R, Kuseyrioglu B, Uzun H, Uludag S, Ocak V. Prediction of preeclampsia with maternal mid-trimester placental growth factor, activin A, fibronectin and uterine artery Doppler velocimetry. Int J Gynaecol Obstet 2005;89:251-7. |
| 16. | Nagtegaal MJ, van Rijswijk S, McGavin S, Dekker G. Use of uterine Doppler in an australian level II maternity hospital. Aust N Z J Obstet Gynaecol 2005;45:424-9. |
| 17. | Rampello S, Frigerio L, Ricci E, Rota E, Lucianetti M, Parazzini F, et al. Transabdominal uterine arteries Doppler at 12-14 th and 20-24 th week of gestation and pregnancy outcome: A prospective study. Eur J Obstet Gynecol Reprod Biol 2009;147:135-8. |
| 18. | Parra M, Rodrigo R, Barja P, Bosco C, Fernández V, Muñoz H, et al. Screening test for preeclampsia through assessment of uteroplacental blood flow and biochemical markers of oxidative stress and endothelial dysfunction. Am J Obstet Gynecol 2005;193:1486-91. |
| 19. | Cooley SM, Donnelly JC, Walsh T, MacMahon C, Gillan J, Geary MP, et al. The impact of umbilical and uterine artery Doppler indices on antenatal course, labor and delivery in a low-risk primigravid population. J Perinat Med 2011;39:143-9. |
| 20. | Noor H, Manzoor M, Sadiq S, Shahzad N. Maternal history and second trimester uterine artery Doppler in the assessment of risk for development of early and late onset pre-eclampsia and intra uterine growth restriction. Int J Reprod Contracept Obstet Gynecol 2016;5:1571-6. |
| 21. | Asnafi N, Hajian K. Mid-trimester uterine artery Doppler ultrasound as a predictor of adverse obstetric outcome in high-risk pregnancy. Taiwan J Obstet Gynecol 2011;50:29-32. |
| 22. | Audibert F, Benchimol Y, Benattar C, Champagne C, Frydman R. Prediction of preeclampsia or intrauterine growth restriction by second trimester serum screening and uterine Doppler velocimetry. Fetal Diagn Ther 2005;20:48-53. |
| 23. | Espinoza J, Kusanovic JP, Bahado-Singh R, Gervasi MT, Romero R, Lee W, et al. Should bilateral uterine artery notching be used in the risk assessment for preeclampsia, small-for-gestational-age, and gestational hypertension? J Ultrasound Med 2010;29:1103-15. |
| 24. | Uludag S, Madazli R, Benian A, Ocak V, Erez S. The relationship between maternal serum alfa fetoprotein and uterine artery Doppler findings at 20-24 weeks' gestation for prediction of preeclampsia and intrauterine growth retardation. Marmara Med J 2002;15:169-74. |
| 25. | Marchesoni D, Pezzani I, Springolo F, Ianni A, Casarsa S, Zavarise D, Driul L. The use of uterine artery Doppler as a screening test for pre-eclampsia. Ital J GynaecolObstet. 2003; 15:15–20. |
| 26. | Aquilina J, Barnett A, Thompson O, Harrington K. Comprehensive analysis of uterine artery flow velocity waveforms for the prediction of pre-eclampsia. Ultrasound Obstet Gynecol 2000;16:163-70. |
| 27. | Albaiges G, Missfelder-Lobos H, Parra M, Lees C, Cooper D, Nicolaides KH. Comparison of color Doppler uterine artery indices in a population at high risk for adverse outcome at 24 weeks' gestation. Ultrasound Obstet Gynecol. 2003; 21:170-3. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7]
| This article has been cited by | | 1 |
Pre-eclampsia in a Sub-Saharan African country and maternal-perinatal outcome |
|
| Elie Nkwabong, Franck Djientcheu Deugoue, Jeanne Fouedjio | | Tropical Doctor. 2022; : 0049475522 | | [Pubmed] | [DOI] | | | 2 |
Utility of uterine artery doppler indices for prediction of preeclampsia: A narrative review with systematic analysis |
|
| PrernaAnup Patwa, GauravVedprakash Mishra, RajasbalaPradeep Dhande, RohanKumar Singh, Sonal Singh | | Journal of Datta Meghe Institute of Medical Sciences University. 2022; 17(3): 757 | | [Pubmed] | [DOI] | | | 3 |
circCRAMP1L is a novel biomarker of preeclampsia risk and may play a role in preeclampsia pathogenesis via regulation of the MSP/RON axis in trophoblasts |
|
| Yonggang Zhang,Hongling Yang,Yipeng Zhang,Junzhu Shi,Ronggui Chen | | BMC Pregnancy and Childbirth. 2020; 20(1) | | [Pubmed] | [DOI] | |
|
 |
|
|
|
Search Pubmed for