Les complications du travail restent les facteurs de risque les plus importants de mortalité périnatale dans les zones rurales du Kenya
Las complicaciones del parto siguen siendo el factor de riesgo de mortalidad perinatal más importante en la Kenya rural
Renay WeinerI, 1; Carine RonsmansII; Ed DormanIII; Hilton JiloIV; Anne MuhoroV; Caroline ShulmanVI
ILecturer and Specialist in Public Health Medicine, School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, 7 York Road, Parktown, 2193, Johannesburg, South Africa (email: firstname.lastname@example.org)
IISenior Lecturer, Infectious Disease Epidemiology Unit, London School of Hygiene and Tropical Medicine (LSHTM), London, England
IIIConsultant Obstetrician, Homerton Hospital, London, England; and Kenya Medical Research Institute, Centre for Geographical Medicine Research, Kilifi, Kenya
IVObstetrician and Gynaecologist, Kilifi District Hospital, Kilifi, Kenya
VMidwife, Kenya Medical Research Institute, Centre for Geographical Medicine Research, Kilifi, Kenya
VIHonorary Senior Lecturer, Department of Infectious and Tropical Diseases (LSHTM), London, England; and Kenya Medical Research Institute, Centre for Geographical Medicine Research, Kilifi, Kenya
OBJECTIVES: To identify and quantify risk factors for perinatal mortality in a Kenyan district hospital and to assess the proportion of perinatal deaths attributable to labour complications, maternal undernutrition, malaria, anaemia and human immunodeficiency virus (HIV).
METHODS: A cross-sectional study of 910 births was conducted between January 1996 and July 1997 and risk factors for perinatal mortality were analysed.
FINDINGS: The perinatal mortality rate was 118 per 1000 births. Complications of labour such as haemorrhage, premature rupture of membranes/premature labour, and obstructed labour/ malpresentation increased the risk of death between 8- and 62-fold, and 53% of all perinatal deaths were attributable to labour complications. Placental malaria and maternal HIV, on the other hand, were not associated with perinatal mortality.
CONCLUSIONS: Greater attention needs to be given to the quality of obstetric care provided in the rural district-hospital setting.
Keywords: Infant mortality; Labor complications/mortality; Maternal nutrition; Prenatal care; HIV infections; Risk factors; Rural population; Cross-sectional studies; Kenya (source: MeSH, NLM).
OBJECTIF: Identifier et dénombrer les facteurs de risque de mortalité périnatale dans un hôpital de district au Kenya et évaluer la proportion de décès périnatals due aux complications du travail, à la dénutrition maternelle, au paludisme, à l'anémie et au virus de l'immunodéficience humaine (VIH).
MÉTHODES: Une étude transversale sur 910 naissances a été réalisée entre janvier 1996 et juillet 1997, et les facteurs de risque de mortalité périnatale analysés.
RÉSULTATS: Le taux de mortalité périnatale était de 118 pour 1000 naissances vivantes. Les complications du travail telles que hémorragie, rupture prématurée des membranes/accouchement prématuré, dystocie/présentation vicieuse, multipliaient le risque de décès de 8 à 62 fois, et 53 % de l'ensemble des décès maternels étaient imputables aux complications du travail. Par ailleurs le paludisme placentaire et le VIH de la mère n'étaient pas associés à la mortalité maternelle.
CONCLUSION: Il convient d'accorder une attention plus grande à la qualité des soins obstétricaux dispensés dans un hôpital de district rural.
Mots clés: Mortalité nourrisson; Accouchement compliqué/mortalité; Nutrition maternelle; Soins prénataux; HIV, Infection; Facteur risque; Population rurale; Etude section efficace; Kenya (source: MeSH, INSERM).
OBJETIVO: Identificar y cuantificar los factores de riesgo de mortalidad perinatal en un hospital de distrito de Kenya y evaluar la proporción de defunciones perinatales atribuibles a complicaciones del parto, desnutrición materna, malaria, anemia y virus de la inmunodeficiencia humana (VIH).
MÉTODOS: Entre enero de 1996 y julio de 1997 se realizó un estudio transversal de 910 nacimientos para analizar los factores de riesgo de defunción perinatal.
RESULTADOS: La tasa de mortalidad perinatal fue de 118 por 1000 nacimientos. Las complicaciones del parto consistentes en hemorragias, ruptura prematura de membranas/parto prematuro, y parto obstruido/presentación defectuosa multiplicaban el riesgo de defunción por un factor de entre 8 y 62, y el 53% de todas las defunciones perinatales se atribuyeron a complicaciones del parto. La malaria placentaria y el VIH materno, en cambio, no se asociaron a mortalidad perinatal.
CONCLUSIÓN: Es necesario prestar más atención a la calidad de la atención obstétrica dispensada en el entorno de los hospitales de distrito rurales.
Palabras clave: Mortalidad infantil; Complicaciones del trabajo de parto/mortalidad; Nutrición materna; Atención prenatal; Infecciones por VIH; Factores de riesgo; Población rural; Estudios transversales; Kenya (fuente: DeCS, BIREME).
Perinatal mortality is an important indicator of obstetric care, health status and socio-economic development (1, 2). Perinatal mortality rates are highest in developing countries, particularly in Africa. In 1995, WHO estimated a perinatal mortality rate of 75 per 1000 births in Africa, a modest decline from the rate of 81 per 1000 births in 1983 and substantially higher than in more-developed countries, where the estimated rate was 11 per 1000 births (1).
The approach to improving maternal and perinatal health in developing countries has shifted in the last decade. In the 1980s, WHO promoted the risk approach which involved screening and risk classification of pregnancies based on maternal characteristics (3). Evidence emerged, however, that complications around the time of labour or delivery and perinatal deaths are not easily predictable, and that antenatal risk screening might be limited in its capacity to reduce maternal and perinatal mortality (4, 5). Consequently, safe motherhood programmes have focused on improving care during labour including strengthening emergency obstetric services (6). Having a health worker with midwifery skills present at delivery is now seen as one of the most critical interventions for making motherhood safer (6).
The extent to which the skilled-attendant approach will also reduce perinatal mortality is less certain. Few studies have simultaneously considered the importance of intra-par-tum morbidity, sociodemographic factors, and prevalent maternal illnesses such as infection with human immunodeficiency virus (HIV), malaria, and poor nutrition as risk factors for perinatal mortality. The proportion of perinatal deaths attributable to such factors also is not well known. The aim of the current study is to identify and quantify risk factors for perinatal mortality in a rural African hospital and to assess the role of labour complications in addition to sociodemographic factors and maternal illnesses such as malaria, HIV infection, anaemia and undernutrition.
Materials and methods
Study sample and population
The study population consisted of women who delivered at Kilifi district hospital, Kenya between January 1996 and July 1997. Kilifi, situated 60 km north of Mombasa on the Kenyan coast, is a predominantly rural, farming district with a population of approximately 700 000 (7). The majority of women attend antenatal care, either at a dispensary, health centre or hospital. Less than half of all women from this district deliver in hospital, so the women are not likely to be representative of the entire study population. Rather, they are a self-selecting group, and in this study area appear to consist of a higher proportion of the more-educated urban women and also a higher proportion of those experiencing complicated deliveries (C. Shulman, personal communication). The study sample comprised approximately 50% of deliveries at the hospital during the study period (8). Since this study was part of a larger one investigating the association between malaria, anaemia, birthweight, pre-eclampsia and intra-uterine fetal deaths, women with any of these complications were more likely to be included (8). Women with multiple pregnancies, significant antenatal haemorrhage prior to admission, and those participating in an antimalarial intervention trial for primigravidae, on the other hand, were excluded.
A combination of data collection methods was used: blood samples were collected on admission for the measurement of blood haemoglobin levels using a Coulter Counter and HIV antibodies using an immunoglobulin G antibody capture particle-adherence test (GACPAT) (9). Placental samples were taken at delivery for histological assessment of malaria. The presence of malaria parasites in erythrocytes in the intervillous space was taken to indicate the presence of active infection and the presence of malaria pigment in fibrin, chronic or past infection (10). A structured questionnaire was administered to each woman after delivery to collect information on sociodemographic markers, antenatal attendance, mode of delivery, birth outcome and condition of her baby at the time of discharge. Maternal weight and height were measured after delivery, just prior to discharge from hospital, and expressed as body mass index (BMI weight (kg)/ height (m) 2). Medical records of discharge diagnoses were reviewed for data on labour complications and we relied on case definitions as noted in the records. Only complications during labour were included. Women with serious antepartum haemorrhage prior to the onset of labour were excluded. Antepartum haemorrhage was defined as a woman who had a discharge diagnosis of antepartum haemorrhage and/or abruptio placentae, but excluded women with placenta previa. Obstructed labour was defined when the medical record documented a diagnosis of either obstructed labour or prolonged labour. Informed consent was obtained from all women prior to participation in the study.
Data were analysed using Stata 7 software. Perinatal death was defined as stillbirths and deaths occurring in the first week of life. Categorical variables were grouped according to risk categories and continuous variables were converted to categorical variables, using WHO criteria for anaemia and quartiles for BMI. Malaria in pregnancy was defined as the presence of histological evidence of active, chronic or past placental malaria infection (10). Associations with perinatal mortality were expressed as odds ratios (OR) with their 95% confidence intervals (CI). Multivariate logistic regression was conducted to assess the strength of the association between antenatal and intra-partum risk factors for perinatal death, while adjusting for all other factors. The independence of each variable and interactions between malaria and anaemia, and malaria and gravidity were assessed using the likelihood ratio test. The population-attributable fraction (PAF) was calculated for the variables included in the final model, based on the adjusted ORs and the prevalence of the risk factor among the exposed (11).
The sample size consisted of 910 births. The mean age of the women was 23.9 years (range 1443 years), approximately 30% were primigravidae, 25% had not received any formal education, and all women for whom there were data attended antenatal clinics at least once (Table 1). Severe anaemia (Hb < 7 g/dl) was present in 14% of the women, 9% were HIV-positive, and 47% had placental malaria. Labour complications were noted for 20% of women, while seven suffered from more than one complication. Obstructed or prolonged labour was the most commonly noted labour complication (8.5%).
There were 108 perinatal deaths giving an overall perinatal mortality rate (PNMR) of 118 per 1000 births. The PNMR was significantly higher among women without any schooling, those with a history of previous stillbirths, and those who had attended fewer than three antenatal visits (Table 1). Perinatal mortality was associated with severe anaemia (OR = 2.2; CI 1.14.3) and low BMI (OR = 3.1; CI 1.75.6), but not with HIV infection status (OR = 0.96; CI 0.461.99) or placental malaria (OR = 0.68, 95% CI 0.441.04). The PNMR for women who had no reported labour complications was 59 per 1000 births compared to 367 per 1000 for women with abnormal labour (OR = 9.31; CI 5.8314.88). Haemorrhage was associated with the highest perinatal mortality rate (750 per 1000 births), followed by malpresentation (714 per 1000 births), and eclampsia (615 per 1000 births). The perinatal mortality rate associated with induced births was extremely high (618 per 1000 births) because stillbirths were an indication for induction of labour rather than vice versa.
The results of multivariate analysis are presented in Table 2. Age, ownership of a latrine and sex of the baby were excluded from the model because they were neither confounders nor independent risk factors for perinatal mortality. Mode of delivery was also excluded because it was on the causal pathway between labour complications and perinatal mortality.
After adjustment for other risk factors and potential confounders, labour complications were the main risk factors for perinatal mortality. Women with an antepartum bleed were over 60 times more likely to have a perinatal death than women without any bleeding (adjusted OR = 61.9; CI 13.9274.2), and the risk of perinatal mortality was eight and 13 times higher, respectively, for women with obstructed or prolonged labour /malpresentation (OR = 7.9; CI 3.9215.94) and premature rupture of membranes/premature labour (OR = 13.6; CI 5.235.7). After adjustment, the risk associated with other complications increased substantially (OR = 52; CI 11.2236.7). When the data were reanalysed excluding labour complications of unknown cause (missing data), a lower risk was measured for the category other (OR = 21.6; CI 2.7167).
Antenatal risk factors that remained significant after adjustment were poor nutrition (OR = 3.6; CI 1.68.1) and previous stillbirths (OR = 2.5; CI 1.35.1). No evidence for significant interactions was found between malaria and HIV infection status (P = 0.49), malaria and anaemia (P = 0.55), malaria and gravidity (P = 0.93), and anaemia and HIV infection status (P = 0.52). The PAF was 53% for labour complications, followed by 25% and 17% for poor nutrition and prior stillbirths, respectively.
This hospital-based study in rural Kenya has shown that labour complications have a very strong effect on perinatal mortality. Complications such as antepartum haemorrhage, obstruction or prolonged/malpresentation, eclampsia, prematurity and premature rupture of membranes increased the risk of perinatal death between 8- and 62-fold, and 53% of all perinatal deaths were attributable to labour complications. Placental malaria and maternal HIV infection status, on the other hand, were not associated with perinatal mortality in this study.
The sample was hospital based and the sampling procedure was not random, hence caution is required in the interpretation of results, particularly attributable fractions. The PNMR, for example, might have been over-estimated, since low-birth-weight babies and intrauterine deaths were oversampled. Conversely, babies who were discharged and died at home within the first week of life may have resulted in not all deaths from the sample being included. Some obstetric complications may have been overrepresented, and providers may have been more likely to record a problem when the baby was stillborn, thus inflating the PAF for labour complications. However, the large odds ratios for the association between labour complications and perinatal mortality cannot be explained on the basis of bias alone. Despite some methodological weaknesses, there is no doubt that labour complications are important causes of perinatal death in this hospital.
Our findings are consistent with those from other studies in Kenya and elsewhere. Kavoo-Linge & Rogo (12) identified prolonged labour as a particularly important factor for perinatal deaths occurring within the first 24 h after hospital delivery in Kenya, while labour complications were associated with perinatal death in almost 40% of deliveries in another rural district. (12, 13). Data from West Africa, Bangladesh, and Guatemala also confirm high perinatal mortality following prolonged labour or malpresentation (1417). Premature labour is a known risk factor for perinatal death (18) and the importance of haemorrhage during labour as a risk factor for perinatal mortality has previously been documented in West Africa and India (14, 19, 20). Although effective interventions may not fully eliminate the risk of death, emergency management of abruptio placentae, early detection and management of malpresentation, and better monitoring of labour could result in significant reductions in perinatal mortality.
Poor maternal nutrition has been associated with increased perinatal mortality in other developing countries (16, 19, 21). Moreover, weight gain during pregnancy appears to be important and decreases in perinatal deaths associated with increasing maternal weight have been demonstrated suggesting that nutritional interventions during pregnancy are important (2123).
Although not statistically significant after adjustment, the risks associated with few antenatal visits need to be interpreted with caution. The large proportion of missing data and recall bias may have further influenced the validity of the results. The content, number and timing of antenatal visits in developing countries has recently received rigorous evaluation. A multi-centred randomized controlled trial comparing the existing versus a revised model of fewer, goal-oriented activities has shown similar outcomes in the two groups (24, 25). This is supported by evidence from Zimbabwe, which has shown that fewer antenatal visits, four versus seven, does not lead to adverse effects on pregnancy (26).
Maternal anaemia and malaria
The lack of a statistically significant association with severe maternal anaemia after adjustment for confounders and the apparently protective effect of malaria on perinatal mortality highlight the complexity of the interrelation between maternal anaemia, placental malaria and gravidity (8). The study may have lacked power to detect interactions between these factors, in particular between primigravidity, anaemia and malaria. Some authors have found an association between maternal anaemia and perinatal mortality (19, 25) while others have not (16). A systematic review of randomized controlled trials has revealed that the existing data on the effects of iron and folate supplementation on perinatal outcomes are very limited and clear benefits have not been demonstrated (27). However, this is unsurprising as most of these trials were conducted in industrialized countries and excluded women with anything other than mild anaemia. Malaria in pregnancy is generally accepted to increase the risk of low birth weight and maternal anaemia (28, 29). In highly endemic areas, however, no clear contribution of placental malaria to perinatal death has been observed (30) and in rural Malawi, placental malaria diagnosed on blood smear appeared to be associated with lower perinatal mortality among normal birthweight babies (31). Further research on the interaction between malaria and anaemia and their impact on perinatal outcomes is required.
Maternal HIV infection status
In this study, maternal HIV infection status was not a risk factor for perinatal deaths, a finding consistent with data from a larger sample in rural Malawi (32). In a meta-analysis of the association between maternal HIV infection and perinatal outcomes a summary odds ratio of 1.79 (CI 1.142.81) was measured for the effect of HIV infection status on perinatal mortality, but publication bias and uncontrolled confounding, particularly in studies from developing countries, may account for part of these findings (33). In Kenya, an association between HIV infection status and intrauterine and intrapartum deaths was found after confounders were controlled (34). The lack of consistency in the associations between HIV infection status and stillbirth rates may reflect the extent of the epidemic and the nature of the HIV-related disease in different communities (35).
Consistent with the limited data available in Africa, the findings of this study indicate that in a hospital setting, labour complications pose the greatest risk for perinatal mortality. Even where labour care is provided, high risks are associated with complications such as obstructed labour and haemorrhage. This presents an opportunity for effective health service intervention. If detected early and appropriately managed, many problems that arise in labour can be limited, and serious morbidity and mortality can be averted. The findings add support to strategies that focus on improving the quality of labour care, and mutual gains in perinatal and maternal health can be expected. However, the role of early detection and management of antenatal risk factors such as hypertension and poor maternal nutrition continues to be important.
This paper has been published with the permission of the Director of the Kenya Medical Research Institute (KEMRI). The original study was funded by the UK Department for International Development (DFID). Particular thanks go to the KEMRI fieldworkers, hospital midwives and the women who participated in the study.
Conflicts of interest: none declared.
1. Perinatal mortality: a listing of available information. Geneva: World Health Organization; 1996. WHO document WHO/FRH/MSM/96.7.
2. Bakketeig LS, Hoffman HJ, Oakley ART. Perinatal mortality. In: Bracken M, editor. Perinatal epidemiology. Oxford: Oxford University Press; 1984. p. 111-20.
3. Backet EM, Davies AM, Petros-Barvazian A. The risk approach in health care. With special reference to maternal and child health, including family planning. Public Health Papers No.76. Geneva: World Health Organization; 1984.
4. Rooney C. Antenatal care and maternal health. How effective is it? A review of the evidence. Geneva: World Health Organization; 1992. WHO document WHO/MSM/92.4.
5. Alexander A, Keirse MJNC. Formal risk scoring during pregnancy. In: Chlamers I, Enkin M, Keirse MJNC, editors. Effective care in pregnancy and childbirth. Oxford: Oxford University Press; 1991. p.345-65.
6. Starrs A. The safe motherhood action agenda: priorities for the next decade. New York: Inter-Agency Group for Safe Motherhood and Family Care International; 1997.
7. Kirumbi LW, Wassuna A, Anabwani G, Mpanju-Shumbsho. Prevention of perinatal and neonatal morbidity and mortality in the East, Central and Southern African Region. Prevalence of and factors associated with perinatal morbidity and mortality in Kirinyaga and Kilifi district in Kenya. Phase 1. June 1996 (unpublished report).
8. Shulman CE, Marshall T, Dorman EK, Bulmer JN, Cutts F, Peshu N, et al. Malaria in pregnancy: adverse effects on haemoglobin levels and birthweight in primigravidae and multigravidae. Tropical Medicine and International Health 2001;6:1-9.
9. Parry JV, Mortimer PP. An immunoglobulin G antibody capture particle-adherence test (GACPAT) for antibody to HIV-1 and HTLV-1 that allows economical large scale screening. AIDS 1989;3:173-6.
10. Bulmer JN, Rasheed FN, Francis N, Morrison L, Greenwood BM Placental malaria. I. Pathological classification. Histopathology 1993;22:211-8.
11. Bruzzi P, Green SB, Byar DP, Brinton LA, Schairer C. Estimating the population attributable risk for multiple risk factors using case-control data. American Journal of Epidemiology 1985;122:904-14.
12. Kavoo-Linge, Rogo KO. Factors influencing early perinatal mortality in a rural district hospital. East African Medical Journal 1992;69:181-7.
13. Voorhoeve AM, Nordbeck HJ, Anderson J, Van Ginneken. Perinatal mortality and the high risk approach in antenatal screening in a rural area in Kenya. East African Medical Journal 1993;60:626-35.
14. Chalumeau M, Salanave B, Bouvier-Colle MH, de Bernis L, Prual A, Breart G. Risk factors for perinatal mortality in West Africa: a population-based study of 20326 pregnancies. Acta Paediatrica 2000;89:1115-21.
15. Wessel H, Cnattingius S, Dupret A, Reitmaier P, Bergstrom S. Risk factors for perinatal death in Cape Verde. Paediatric Perinatal Epidemiology 1998;12:25-36.
16. Kusiako T, Ronsmans C, Van der Paal L. Perinatal mortality attributable to complications of childbirth in Matlab, Bangladesh. Bulletin of the World Health Organization 2000;87:621-7.
17. Bartlett A, Paz de Bocaletti ME, Bocaletti MA. Reducing perinatal mortality in developing countries: high risk or improved labour management? Health Policy and Planning 1993;8:360-8.
18. Villar J, Ezcurra EJ, Gurtner de la Fuente, Campodonico L. Pre-term delivery syndrome: the unmet need. In: Keirse MJNC, editor. New perspectives for the effective treatment of pre-term labour. An international consensus. Research & Clinical Forum, Weld Medical, England; 1994;16:9-33.
19. Mavalankar DV, Trivedi C, Gray RH. Levels and risk factors for perinatal mortality in Ahmedabad, India. Bulletin of the World Health Organization 1991;69:435-42.
20. Omene JA, Diejomaoh FME. Factors influencing perinatal mortality in a Nigerian community. East African Medical Journal 1977;54:202-6.
21. Barros FC, Victora CG, Vaughan JP, Estanislau HJ. Perinatal mortality in southern Brazil: a population-based study of 7392 births. Bulletin of the World Health Organization 1987;65:95-105.
22. Neumann C, Ferguson L, Bwibo NO. Maternal anthropometry as a risk predictor of pregnancy outcome: the Nutrition CRSP in Kenya. Bulletin of the World Health Organization 1995;73:91-5.
23. Osman NB, Challis K, Cotiro M, Nordahl G, Bergstrom S. Perinatal outcome in an obstetric cohort in Mozambican women. Journal of Tropical Paediatrics 2001;47:30-8.
24. Villar J, Ba'aqeel H, Piaggio G, Lumbiganon P, Miguel Belizan J, Farnot U, et al. WHO antenatal care randomised trial for the evaluation of a new model of routine antenatal care. Lancet 2001;19:1551-64.
25. Villar J, Bakketeig L, Donner A, Al-Mazrou Y, Ba'aqeel H, Belzian JM, et al. The WHO antenatal care randomised controlled trial: rationale and study design. Paediatric and Perinatal Epidemiology 1998;12 Suppl 2:27-58.
26. Manjanja SP, Lindmark G, Nystrom L. Randomised controlled trial of reduced-visits programme of antenatal care in Harare, Zimbabwe. Lancet 1996;348:364-9.
27. Mohamed K. Iron and folate supplementation in pregnancy (Cochrane Review). In: The Cochrane Library, Issue 3. Oxford: Update Software; 2001.
28. Brabin BJ. The risks and severity of malaria in pregnant women. Applied Field Research in Malaria Reports. Geneva: World Health Organization; 1991. WHO
29. Steketee RW, Wirima JJ, Hightower AW, Slutsker L, Heymann DL, Breman JG. The effect of malaria and malaria prevention in pregnancy on offspring birth weight, prematurity, and intrauterine growth retardation in rural Malawi. American Journal of Tropical Medicine and Hygiene 1996;55 Suppl 1:33-41.
30. Steketee RW, Wirima JJ, Slutsker L, Heymann DL, Breman JG. The problem of malaria control in pregnancy in sub-Saharan Africa. American Journal of Tropical Medicine and Hygiene 1996;55:2-7.
31. McDermott J, Wirima JJ, Steketee RW, Breman JG, Heymann DL. The effect of placental malaria infection on perinatal mortality in rural Malawi. American Journal of Tropical Medicine and Hygiene 1996;55:61-5.
32. McDermott J, Steketee R, Wirima J. Perinatal mortality in Malawi. Bulletin of the World Health Organization 1996;74:165-71.
33. Brockelhurst P, French R. The association between maternal HIV infection and perinatal outcome: a systematic review of the literature and meta-analysis. British Journal of Obstetrics and Gynaecology 1998;105:836-48.
34. Timmerman M, Plummer FA, Mirza NB. Infection with HIV as a risk factor for adverse pregnancy outcome. AIDS 1990;4:139-44.
35. McIntyre J. HIV in pregnancy: a review. Geneva: World Health Organization Joint United Nations Programme on HIV/AIDS; 1998. WHO document WHO/CHS/RHR/99.15.
1 Correspondence should be addressed to this author.