Corresponding Author Patrick F. van Rheenen, Department of Paediatrics, Paediatric Gastroenterology, University Medical Center Groningen, University of Groningen, PO Box 30001, 9700 RB Groningen, The Netherlands. Tel.: +31 50 3614151; Fax: +31 50 3612742; E-mail: firstname.lastname@example.org
Objectives This study was carried out to assess whether delaying umbilical cord clamping is effective in improving the haematological status of term infants living in a malaria-endemic area, and whether this is associated with complications in infants and mothers.
Methods We randomly assigned women delivering term babies in Mpongwe Mission Hospital, Zambia, to delayed cord clamping (DCC, n = 46) or immediate cord clamping (controls, n = 45) and followed their infants on a bi-monthly basis until the age of 6 months. We compared the haemoglobin (Hb) change from cord values and the proportion of anaemic infants. Secondary outcomes related to infant and maternal safety.
Results Throughout the observation period infant Hb levels in both groups declined, but more rapidly in controls than in the DCC group [difference in Hb change from baseline at 4 months 1.1 g/dl, 95% confidence interval (CI) 0.2; 2.1]. By 6 months, this difference had disappeared (0.0 g/dl, 95% CI −0.9; 0.8). The odds ratio for iron deficiency anaemia in the DCC group at 4 months was 0.3 (95% CI 0.1; 1.0), but no differences were found between the groups at 6 months. No adverse events were seen in infants and mothers.
Conclusion Our findings indicate that DCC could help improve the haematological status of term infants living in a malaria-endemic region at 4 months of age. However, the beneficial haematological effect disappeared by 6 months. This simple, free and safe delivery procedure might offer a strategy to reduce early infant anaemia risk, when other interventions are not yet feasible.
Objectifs Evaluer si le retardement la coupure du cordon ombilical est efficace pour améliorer le statut hématologique des enfants en bas âge nés à terme, vivant dans une zone endémique pour la malaria et voir si cela est associéà des complications chez les enfants en bas âge et les mères.
Méthodes Nous avons aléatoirement appliqué aux femmes donnant naissance à terme dans l'hôpital missionnaire de Mpongwe en Zambie, soit le retardement de la coupure du cordon (n = 46), soit sa coupure immédiate (contrôles, n = 45) et avons suivi deux fois par mois leurs enfants en bas âge jusqu’à l’âge de six mois. Nous avons comparé la variation de l'hémoglobine (Hb) à partir des valeurs initiales du cordon et la proportion d'enfants en bas âge anémiques. Les résultats secondaires concernaient la sûreté infantile et maternelle.
Résultats Tout au long de la période d'observation les taux d'hémoglobine infantile dans les deux groupes ont diminués, mais plus rapidement chez les contrôles que dans le groupe avec retardement de la coupure du cordon (différence de variation de Hb à partir de la ligne de base à quatre mois 1,1 g/dl, intervalle de confiance 95% (IC95%): 0,2–2,1). A 6 mois, cett e différence avait disparu (0,0 g/dl, IC95%: −0,9–0,8). Les rapports de cote (OR) pour l'anémie par insuffisance de fer dans le groupe avec coupure retardée du cordon à quatre mois était 0,3 (IC95%: 0,1–1,0), mais aucune différence n'a été trouvée entre les deux groupes à six mois. Aucun événement adverse n'a été observé chez les enfants en bas âge et les mères.
Conclusion Nos résultats indiquent que le retardement de la coupure du cordon pourrait aider à améliorer le statut hématologique des enfants en bas âge nés à terme et vivant dans une région endémique pour la malaria à quatre mois d’âge. Cependant, l'effet bénéfique hématologique a disparu à six mois. Ce procédé simple, sans frais et sûr de l'accouchement pourrait offrir une stratégie pour réduire le risque d'anémie infantile initiale, quand d'autres interventions ne sont pas encore réalisables.
Objetivos Evaluar si el retrasar el pinzamiento del cordón umbilical es efectivo para mejorar el estado hematológico de bebés nacidos a término en áreas endémicas para malaria, y si existe alguna asociación con complicaciones en los niños y en las madres.
Métodos Asignamos de forma aleatoria a mujeres que se encontraban en trabajo de parto en el Hospital Mpongwe Mission, en Zambia, a un pinzamiento tardío (PT, n = 46) o a un pinzamiento inmediato (controles, n = 45), y se realizó un seguimiento a los bebés cada dos meses y hasta que cumplieron los seis meses. Comparamos los cambios en hemoglobina a partir de los valores obtenidos en el cordón umbilical y la proporción de neonatos anémicos. Los resultados secundarios estaban relacionados con la seguridad materna e infantil.
Resultados A lo largo del periodo de observación los niveles de hemoglobina de ambos grupos de neonatos bajaron, pero más rápidamente en los controles que en el grupo de PT (diferencia en cambio de Hb de la línea de base a los cuatro meses 1.1 g/dl, 95% intervalo de confianza (IC) 0.2; 2.1). A los 6 meses, esta diferencia había desaparecido (0.0 g/dl, 95% CI −0.9; 0.8). La razón de probabilidades para la anemia por deficiencia de hierro en el grupo PT a los cuatro meses era de 0.3 (95% IC 0.1; 1.0), pero no se hallaron diferencias entre los grupos a los seis meses. No se observaron eventos adversos en niños y neonatos.
Conclusión Nuestros resultados indican que el PT podría ayudar a mejorar el estado hematológico a los cuatro meses de edad, de niños nacidos a término que viven en regiones endémicas para malaria. Sin embargo, el efecto hematológico beneficioso desaparece a los seis meses de edad. Este procedimiento durante el parto, es simple, gratis y seguro, y podría ofrecer una estrategia para reducir el riesgo de anemia temprana en neonatos, cuando otras intervenciones aún no son factibles.
One of the most critical factors contributing to neonatal and infant mortality in developing countries is anaemia (Brabin et al. 2001, 2003b), which has been repeatedly shown to be an intractable problem even with antimalarial and iron interventions. One of the most important trials which studied primary anaemia prevention in Tanzania observed a 20% incidence of severe anaemia even in infants who had received prophylactic antimalarials and iron supplements (Menendez et al. 1997). Although iron supplementation in infants from malarious regions has been shown to be beneficial it remains a highly contentious issue, as it might worsen the outcome of infectious illnesses (Oppenheimer 2001; English & Snow 2006; Sazawal et al. 2006).
In view of this there is interest in evaluating the potential for improving the iron status of infants by enhancing their red cell mass with delayed umbilical cord clamping. In traditional African home deliveries the umbilical cord is cut after placental descent into the vagina (Lefeber & Voorhoeve 1997), but in hospital deliveries in resource-poor settings immediate cord clamping (ICC) is the routine standard of care. It is estimated that the total fetoplacental blood volume is roughly 120 ml/kg of fetal weight. After ICC the distribution of blood reflected in the fetus:placenta ratio is approximately 2:1. Allowing placental transfusion to occur for at least 3 min results in a larger fetal blood volume with 15 ml/kg of blood remaining in the placenta (Yao et al. 1969; Linderkamp 1982). Compared with immediate clamping, a clamping delay of 3 min provides an additional blood volume of 20–35 ml/kg of body weight (Yao et al. 1969; Linderkamp et al. 1992). For a 3-kg infant with a packed cell volume (PCV) of approximately 0.50 at birth, this amounts to an additional 45 mg of iron added to iron stores. Theoretically this amount of iron should be sufficient to meet the requirements of an infant for more than 3 months (Oski 1993).
Previous trials indicated that delayed cord clamping (DCC) prevents a steep decline in haemoglobin (Hb) concentration at 2–3 months of age, especially in infants born to anaemic mothers (van Rheenen & Brabin 2004). A recent trial from Mexico showed that at 6 months of age, infants who had DCC had a better iron status compared with early clamped infants (Chaparro et al. 2006).
It has never been evaluated whether this delivery procedure is effective in reducing anaemia in infants from malaria-endemic regions. We performed a randomised controlled trial in a highly malarious rural area of Zambia to assess whether DCC is effective in reducing anaemia in term infants up to the age of 6 months, and is associated with complications in infants and mothers.
Study area, enrolment and study population
The study location was Mpongwe District, a rural region of the Copperbelt Province in Zambia, approximately 1000 m above sea level. Malaria transmission in this area is holoendemic. Peak malaria transmission occurs from November to April. Like in other parts of Zambia, over 50% of women deliver at home (Stekelenburg et al. 2004). Especially nulliparous and grand multiparous women, and women with a prior caesarean section, are advised to deliver in hospital. Those who are able to pay the user fees and live in the vicinity are more likely to deliver in hospital. The caesarean section rate in Mpongwe Mission Hospital is 5%, with the majority of sectioned women being transferred from peripheral health units for obstructed labour. Approximately 10% of the babies are born with birth weight below 2500 g.
Full-term pregnant women delivering in Mpongwe Mission Hospital were candidates for inclusion in the study. Prespecified exclusion criteria were: (1) twin pregnancy, (2) history of post partum haemorrhage (PPH), (3) gestational diabetes, (4) pre-eclampsia, (5) placental separation before delivery, (6) caesarean section, (7) tight nuchal cord necessitating early cutting, (8) need for neonatal resuscitation or (9) major congenital abnormalities (e.g. neural tube defects). Criteria 1–4 were applied before randomisation. Criteria 5–9 were assessed after randomisation. Infants who weighed <2500 g, or with gestational age below 37 weeks, were excluded.
We investigated in a partially blinded randomised controlled trial: (i) whether DCC affected the haematological status of term infants in a malaria-endemic region and (ii) whether DCC is associated with complications in infants and mothers. Primary endpoints included Hb change from cord values and the proportion of anaemic infants at 4 months after birth, and the duration infants remained free of anaemia during follow-up. Secondary outcomes included possible side effects of DCC in infants (PCV changes 1 day postpartum; clinical signs of hyperviscosity syndrome or hyperbilirubinaemia) and mothers (Hb change 1 day after delivery in relation to antenatal values and blood loss in the third stage of labour). Infants were followed-up every 2 months until 6 months of age.
Initially this trial had a follow-up of 4 months and ‘Hb change from baseline at 4 months’ was one of the prespecified summary measures to analyse serial Hb-values. This summary measure was of clear clinical and biological relevance, as no previous cord clamping trial had done a follow-up beyond 3 months after birth. During the trial we decided to extend the follow-up period from 4 to 6 months, to evaluate how long the potential beneficial effects of DCC would be measurable. The Research Ethics Committee of the Liverpool School of Tropical Medicine and all other parties approved the extension. As the shape of the ‘time-response’ curve was uncertain, we could only choose a new summary measure after the data had been examined, and we decided to stick to the initial summary measures.
Pregnant women were randomised to either DCC or ICC. ICC was the routine standard of care in Mpongwe Mission Hospital at the time of the trial, and was usually completed within 20 s of delivery. In the DCC group the umbilical cord was clamped after the cord stopped pulsating. The cord was palpated every other 20 s for cessation of pulsations and clamped when pulsations were no longer observed. The exact time was recorded by stopwatch.
For allocation concealment, the randomisation instructions were given to study midwives in sequentially numbered, opaque, sealed envelopes with unpredictable allocation code, which were only opened when a mother had consented to enrol. Used envelopes with the assignment instruction enclosed were sent to the study coordinator and were regularly audited. Randomisation was done on admission to the labour ward, when the women were in the first stage of labour and the attending midwife was convinced that uterine contractions would persevere. Although study staff did not inform the mothers of their assignment, the nature of the intervention made it impossible to blind them. If an already randomised mother later became ineligible, the assigned allocation code was not re-used. One of the investigators (De Moor) monitored the delivery procedure, and was therefore not blinded to treatment assignment.
A structured survey questionnaire was used to gather obstetrical and medical details and a venous blood sample was taken from the mother for laboratory investigations in the first stage of labour. After vaginal birth all infants were placed between the legs of the mother (approximately 10 cm below the vaginal introitus), dried and wrapped in a warm towel. The infants remained in this position until the cord was clamped. Intramuscular oxytocin was administered to the mothers after clamping of the cord.
A sample of cord blood was collected for laboratory analysis. Maternal blood loss was estimated visually by the attending midwife. A more objective measure of blood loss was Hb concentration 24 h postpartum compared with Hb in the first stage of labour.
Birthweight was measured using a Salter balance (nearest 10 g). In the first 24 h after delivery the child was observed for clinical signs of hyperviscosity syndrome and hyperbilirubinaemia by one of the investigators, who was not blinded to the assigned intervention. The icterometer, a Perspex ruler with yellow stripes, was used to estimate the degree of hyperbilirubinaemia. In the darkly pigmented newborns in our study population the icterometer was used to blanch the gum. This method has been shown to be as useful as the original (Morley 1973; Hamel 1982). Jaundice was assessed in daylight at a window. We looked for hyperviscosity syndrome whenever the babies’ vital signs were measured, by examining for plethora, apathy, tachypnea, poor sucking and hypoglycaemia. Before discharge (16–24 h after delivery), assessment of gestational age was completed by the Ballard-ext method (Verhoeff et al. 1997); a capillary heel-prick blood sample was collected from the baby after prewarming and a venous sample was taken from the mother.
For mother-infant pairs living within a radius of 4 km from Mpongwe Mission Hospital, follow-up took place at the hospital-based Mother and Child Health Clinic, where vaccinations and growth monitoring were routinely undertaken. Mothers who indicated a preference to attend a more local health facility for vaccination and growth monitoring and nonattendees were traced in their villages by the investigators to minimise loss to follow-up. At 2 months baseline data concerning socio-economic and demographic background were collected to assess possibly confounding factors. Literacy status was assessed by asking the mother to read a simple sentence in the local language. Weight of the mothers was estimated in bare feet (nearest kg) using a standing weigh scale and height (nearest cm) using a height board. Mid-upper-arm circumference (MUAC) was measured on the right arm, hanging loosely, with a TALC insertion tape (TALC, St Albans, UK) and recorded to the nearest 0.1 cm. At each follow-up visit information was collected on infant feeding practices and morbidity in the previous 2 months. Infants’ finger prick blood was collected at 2, 4 and 6 months. Infants with Hb concentrations <7 g/dl at follow-up were given therapeutic iron supplements for 2 months. Infants with malaria were treated with a dose of sulphadoxine-pyrimethamine as per local guidelines. All other illnesses were treated or referred if appropriate.
Blood samples were collected in EDTA microtainers (Becton Dickinson, Franklin Lakes, NJ, USA) and immediately stored at 4 °C. Within 6 h the Hb level was measured using a HemoCue (HemoCue AB, Ängelholm, Sweden). Zinc-protoporphyrin (ZPP) level, in μmol ZPP/mol haem, was assessed using a ZP Haematofluorometer (Aviv Biomedical, Lakewood, NJ, USA) within 48 h of collection, PCV by Micro-Haematocrit and blood glucose level in the first 24 h by glucometer (Ascensia Elite XL; Bayer B.V., Mijdrecht, The Netherlands). Blood smears were examined for malaria parasites. A thick smear was considered negative if 100 microscopic fields revealed no parasites. For positive smears, malaria parasites were counted against 300 leucocytes. The placenta tissue samples were kept up to 9 months until processed and embedded in paraffin wax by standard techniques. Paraffin sections 4 μm thick were stained with haematoxylin and eosin.
Anaemia was defined as a Hb concentration more than two standard deviations (SD) below the mean of similarly aged infants from an iron-supplemented USA reference population not exposed to malaria (Dallman 1988). Reference values for the ages of 2, 4 and 6 months were 9.4, 10.3 and 11.0 g/dl respectively. A cut-off of 10.5 g/dl for 6 months old infants has been shown to be more appropriate, and we therefore considered infants with Hb measurements below this value as anaemic (Domellof et al. 2002). Fetal anaemia was defined as a cord Hb < 12.5 g/dl, which is 2 SD below the mean cord Hb for nonmalarious western populations (Brabin 1992). Anaemia in pregnant women was defined as Hb < 11 g/dl (WHO 2001). Iron deficiency was ZPP above 80 μmol/mol haem for adults and infants (Hinchliffe 1999; Rettmer et al. 1999; Domellof et al. 2002; Juul et al. 2003; Soldin et al. 2003; Kling 2006; Lott et al. 2006). A recently published cross sectional study among babies born to nonanaemic mothers found that infants born after 35 weeks completed weeks of gestation had mean ZPP levels in cord blood of 73 μmol/mol haem (Lott et al. 2006). As this result corresponds to normal values in older infants, we decided to also use 80 μmol/mol haem as cut-off for fetal iron deficiency. The combination of abnormal values for both ZPP and Mean Cell Haemoglobin Concentration (MCHC) is a more sensitive indicator of iron deficiency (INACG 1985). We used MCHC cut-offs of 32 g/dl for pregnant women (Letsky 1991) and 28 g/dl for newborns (Saarinen & Siimes 1978; Serjeant et al. 1980). MCHC estimates were not available at 4 and 6 months follow-up. Iron deficiency anaemia in mothers and newborn infants was defined as the combination of ZPP above the cut-off level and MCHC below the cut-off level, together with Hb more than 2 SD below the reference mean. For infants aged 4 and 6 months the combination of ZPP above the cut-off level and Hb more than 2 SD below the reference mean was used.
Malaria was defined as the presence of asexual stage parasites in thick smears at follow-up, independent of the presence of clinical signs and symptoms. A history of a positive blood smear result at a local health facility together with clinical signs or symptoms of malaria in the last 2 weeks before review, was also considered as evidence for a recent malaria infection.
The histopathological slides were classified as ‘acute infection’ when parasites without pigment deposition were seen, as ‘chronic infection’ when parasites and a significant amount of pigment deposition were seen, and as ‘past infection’ when pigment deposition without parasites were seen (Ismail et al. 2000).
Neonatal tachypnea was defined as a respiratory rate over 60/min, hypoglycaemia as glucose below 45 mg/dl (or 2.5 mmol/l) on the first day after birth, polycythaemia as a PCV above 0.70 in capillary blood. The threshold for phototherapy was set at a bilirubin level above 6 mg/dl (grade 2 of the icterometer) at 12 h after birth (Subcommittee on Hyperbilirubinemia 2004).
Underweight was defined as more than 2 SD below the mean of the 2000 CDC reference curves for weight-for-age (Kuczmarski et al. 2000). Intermittent presumptive treatment with antimalarials was defined as at least two doses of sulphadoxine-pyrimethamine during the second and third trimesters of pregnancy. Adolescence was <20 years of age.
The sample size was determined on the basis of a mean Hb change from cord values of 4 (1 SD) g/dl and a difference in mean Hb change between the DCC and control group at 3 months of 1 g/dl (van Rheenen & Brabin 2004). The 1 g/dl was chosen on the basis of clinical relevance: a survival analysis of Malawian infants indicated that a decrease in Hb by 1 g/dl after 6 months of age increased the risk of dying before 12 months of age by 72% (Brabin et al. 2003a). With a study power of 95% and to detect a significant difference with P = 0.05 (two-sided), 28 babies were required for each study group. Allowing for a maximum drop-out of 40%, we planned to enrol 47 babies per group, or 94 in total. A total of 120 envelopes were prepared for the trial with a 1:1 randomisation ratio. As not all envelopes were expected to be used, we could not ensure that comparison groups were exactly the same size.
Data were collected on standardised forms, and analysed with SPSS for Windows, (version 12.0.1 (2003), SPSS Inc., Chicago, IL, USA) and EPI INFO (version 3.2 (2004), CDC, Atlanta, GA, USA). Student t-tests and chi-square tests were used to compare baseline characteristics between treatment groups, and between infants who completed the study and those who were lost to follow-up. For nonparametric data the Mann–Whitney U-test was used. Time to event data were analysed by Kaplan–Meier and log-rank test. All tests were two-tailed. As part of the exclusion criteria could only be applied after randomisation, which is after delivery, data were analysed according to the per-protocol principle. Infants with tight nuchal cord necessitating early cutting, perceived need for neonatal resuscitation and major congenital abnormalities were excluded regardless of group assignment, thus reducing the likelihood of introducing bias. Furthermore, an intention-to-treat analysis would not have been appropriate for examining adverse effects (Altman et al. 2001).
The haematological follow-up data were analysed as continuous variables with analysis of covariance (ancova) and as categorical variables with logistic regression. Baseline imbalances which could influence infant haematological outcome (maternal Hb, parity, age and cord Hb) were controlled for in these analyses. The level of significance used was a P-value < 0.05. This study is registered with the http://www.controlled-trials.com identifier ISRCTN48735857.
The women were contacted while in their first stage of labour to obtain informed consent. The consent information sheet was written in Lamba, the local language. In case of illiteracy, the midwife on duty read the form. Mothers gave written consent or a thumb impression. The research protocol was approved by the Research Ethics Committee of the Liverpool School of Tropical Medicine and by the Board of Management of Mpongwe Mission Hospital. The Mpongwe District Health Management Team, the chairpersons of the health neighbourhoods (lay people representing the community for health issues) and the local chiefs were informed and all provided signed consent.
Figure 1 shows the trial profile. Between May and July 2004, 105 pregnant women were assessed for eligibility before entering the labour ward and randomised to either DCC (n = 55) or control group (n = 50). Nine cases in the DCC group and four cases in the control group were excluded from the analyses. No women delivered by caesarean section after randomisation 91 mother-infant couples entered the trial and were actively followed-up every 2 months until the infants reached the age of 6 months. The follow-up period ended in January 2005. Thirty-five of 46 mother-infant pairs in the DCC, and 37 of 45 pairs in the control group completed 6 months of follow-up (total drop out rate, respectively, 24% and 18%, not significant).
During the follow-up period four children died: one in the neonatal period, the others after 4 months of age. In two of these infants mild anaemia had been diagnosed during the follow-up visit prior to death (Hb > 7.0 g/dl). Baseline characteristics for women and infants who completed the study did not differ significantly from those lost to follow-up.
Table 1 shows that the mothers in the DCC and control groups were comparable in terms of age, parity, nutritional state, antenatal iron supplementation, the use of intermittent preventive antimalarial treatment, socio-economic status and literacy level. None of them had a prior caesarean delivery. The median number of visits to the antenatal clinic was higher in the DCC group than in the control group (4 vs. 3; P = 0.045). This factor, however, was not associated with outcome. Maternal haematological baseline characteristics and histopathological classification of placenta tissue did not significantly differ.
Table 1. Maternal baseline characteristics
Delayed cord clamping (n = 46)
Controls (n = 45)
Data were mean (SD) unless indicated otherwise.
BMI, body mass index; MUAC, mid upper arm circumference; Hb, haemoglobin; PCV, packed cell volume; MCHC, mean cell haemoglobin concentration; ZPP, zinc-protoporphyrin.
Age (years) – median (range)
Chronic energy deficiency
Last interpartum interval (months) – median (range)
Number of antenatal visits – median (range)
Intermittent presumptive treatment for malaria
Iron supplementation in pregnancy
Proportion maternal anaemia
ZPP (mcmol/mol haem)
Proportion iron deficient erythropoiesis (ZPP > 80 μmol/mol haem)
Histopathological details of placenta
In the DCC group the mean (SD) clamping time was 305 s (136) and in the control group 15 s (8). Although ICC was usually done within 20 s after delivery, in four cases the delay was longer than intended (range 22–45 s). Both groups of infants were comparable at birth in terms of anthropometric parameters. At baseline, mean cord Hb was slightly, but not significantly lower in babies in the DCC group.
Table 2 shows that 18% of newborns in the DCC group had fetal anaemia compared with 7% in the control group (P > 0.05), and more than 70% of the infants had iron-deficient erythropoiesis at birth, indicated by a ZPP value above the cut-off of 80 μmol mol haem. In four of 10 cases of fetal anaemia there was associated maternal anaemia. There was a marginally significant relationship between maternal and cord Hb (r = 0.175, P(one-tailed) = 0.053).
Proportion with iron deficient erythropoiesis (ZPP > 80 μmol/mol haem)
Up to a mean (SD) postnatal age of 16 (8) h both mother and child were assessed for possible side-effects of the intervention. Maternal blood loss in the third stage of labour was comparable in both groups and PPH did not occur. Manual removal of the placenta was not required for any of the women involved. The difference between ante- and postnatal maternal Hb levels was small and comparable in both groups (data not shown).
Table 3 shows that the mean increase in PCV on the first day postpartum was significantly higher in newborns from the DCC group (0.13 vs. 0.07; P < 0.001). The proportion of neonates with polycythaemia (PCV > 0.70) was 5% in each group. None of the infants showed clinical signs of hyperviscosity syndrome or hyperbilirubinaemia before discharge.
Table 3. Infant haematological outcome on first day postpartum
Delayed cord clamping (n = 46)
Controls (n = 45)
Effect size (95% CI)
Data were mean (SD) unless indicated otherwise.
PCV, packed cell volume.
PCV increase compared with cord blood
Proportion above phototherapy threshold
Table 4 shows that infant Hb-levels in both the DCC and control groups continued to decline throughout the observation period. By 4 months the decrease in Hb compared with cord blood values was significantly smaller in the DCC group [mean difference 1.1 g/dl, 95% confidence interval (CI) 0.2–2.1 g/dl]. By 4 months more infants in the control group had Hb levels below −2 SD cut-off level and were classified as having iron deficiency anaemia than those with delayed clamping, but the differences were marginally significant (odds ratio 0.3, 95% CI 0.1; 1.0). At 4 months no infant had an Hb level <7.0 g/dl.
Table 4. Infant haematological status at 4 and 6 months
Analyses adjusted for maternal haemoglobin, age, parity and cord haemoglobin†
Delayed cord clamping
Difference or OR (95% CI)
Delayed cord clamping
Difference or OR (95% CI)
Data were mean (SD) unless indicated otherwise. Effect size is expressed as mean difference for continuous variables and as odds ratio for categorical variables.
†Data were adjusted mean values and SD of unadjusted mean values
Iron deficient erythropoiesis (cut-off > 80 μmol/mol haem)
0.8 (0.3; 2.2)
0.8 (0.3; 2.1)
Proportion iron deficiency anaemia (ZPP > 80 μmol/mol haem and Hb < 10.3)
0.3 (0.1; 1.0)
0.3 (0.1; 1.0)
Proportion with positive malaria smear
Analyses adjusted for maternal haemoglobin, age, parity, cord haemoglobin and infant malaria infection†
−0.3 (−1.2; 0.5)
0.0 (−0.7; 0.7)
Hb change from baseline (g/dl)
0.6 (−0.5; 1.7)
0.0 (−0.9; 0.8)
Proportion anaemia (cut-off 10.5 g/dl)
1.4 (0.3; 2.0)
1.3 (0.4; 4.0)
ZPP (μmol/mol haem)
11 (−43; 65)
−7 (−56; 43)
Iron deficient erythropoiesis (cut-off > 80 μmol/mol haem)
1.5 (0.4; 5.2)
Proportion iron deficiency anaemia (ZPP > 80 μmol/mol haem and Hb < 10.5)
1.2 (0.5; 3.0)
1.5 (0.5; 4.4)
Proportion with positive malaria smear
1.4 (0.3; 5.6)
At 6 months, the difference in Hb change from cord values had disappeared and half of the infants in both groups were anaemic. Three infants were given therapeutic iron supplements as their Hb levels dropped below 7.0 g/dl. Results did not differ when analyses were done either with or without control for maternal baseline differences (anaemia, age and parity). The results also did not change when additional analyses were done by actual clamping time rather than by randomisation group. Malaria infections were not diagnosed before the age of 4 months. At 6 months, malaria was diagnosed evenly in both groups. Controlling for infant malaria infections did not change the results at 6 months. Mean ZPP levels and proportion of infants with iron-deficient erythropoiesis were similar in the allocation groups, at 4 and 6 months.
Table 5 shows that infant feeding patterns were similar in both groups, with early introduction of complementary foods. By 4 months less than a third of the children were exclusively breastfed. The early introduction of complementary foods was associated with a gradual increase in the number of diarrhoea episodes, as reported by the mother. At 6 months more than a quarter of the infants suffered from diarrhoea in the 2 weeks prior to the follow-up visit. Figure 2 shows that in the first 4 months weight gain was above the reference for both DCC and control infants (Kuczmarski et al. 2000). Hereafter a reduction in mean weight increments was observed. Weight-for-age Z scores showed a steep fall between 4 and 6 months.
Table 5. Follow-up dietary data
Delayed cord clamping (n = 43)
Controls (n = 41)
Delayed cord clamping (n = 39)
Controls (n = 39)
Delayed cord clamping (n = 35)
Controls (n = 37)
Proportion exclusively breastfed
Proportion on iron supplements
Proportion reported to have had diarrhoea in previous 2 weeks
In the DCC group seven babies had fetal anaemia, of whom five remained free of anaemia throughout the follow-up period. In the control group three babies were anaemic at birth, and one of them was free of anaemia at 6 months follow-up. All fetal anaemia cases were included in the Kaplan–Meier survival curves (Figure 3). The log-rank test showed that there was no significant difference in the proportion of anaemia cases between the groups.
Delaying clamping of the umbilical cord until the cord pulsations cease improved the haematological status of term infants living in a highly malarious area at 4 months of age, as measured by Hb change from cord blood values and anaemia incidence. This intervention was not associated with increased risk of maternal blood loss or with hyperviscosity syndrome in the baby. However, the beneficial effect of extra red cell mass had disappeared between the age of 4 and 6 months, even when malaria infections were statistically controlled for.
Two previous randomised controlled trials on DCC have reported beneficial effects on infant Hb, but the infants were not followed beyond 3 months of age (Grajeda et al. 1997; Gupta & Ramji 2002). A recently published trial from Mexico had a follow-up of 6 months, but did not find a difference in Hb, although the iron status at 6 months was significantly higher in the DCC group. This lack of difference in Hb is most likely due to the fact that iron deficiency was relatively uncommon in the Mexican study population. Hb is normally not affected until iron stores are depleted (Chaparro et al. 2006). All three trials were carried out in nonmalarious areas. This study is the first to evaluate the haematological effects of DCC in infants living in a malaria-endemic area. We were able to continue following the infants after complementary feeding had commenced, and malaria incidence had increased.
Infant Hb levels in the study population continued to decline after the physiological nadir around 2 months. This pattern is observed in other trials from sub-Saharan Africa and many infants reach their lowest point only 6–12 months after birth (Kitua et al. 1997; le Cessie et al. 2002; van Eijk et al. 2002). Anaemia is usually multi-factorial in origin, but it is clear that malaria plays a key aetiological role in endemic countries. In holoendemic malarious areas, infants are infected with malaria from birth onwards but clinical malaria usually only develops from about 4 months of age, when immunity acquired from the mother is wearing off. In these regions the highest burden of anaemia occurs towards the end of the first year of life (Schellenberg et al. 2003).
The disappearance of the haematological benefits of DCC after 4 months can possibly be explained by two pathophysiological mechanisms. Either the infants outgrow their iron stores, or the infants suffer increased loss of iron via the gastrointestinal tract during episodes of diarrhoea or during the feeding of whole cow's milk (Oski 1993).
Neonatal polycythaemia, hyperviscosity and hyperbilirubinaemia are potential consequences of placental transfusion (Behrman et al. 2000), but the safety of DCC in appropriate-for-gestational-age term infants has been demonstrated in several trials (Linderkamp et al. 1992; Nelle et al. 1993, 1995, 1996; Grajeda et al. 1997; Emhamed et al. 2004; van Rheenen & Brabin 2004). However, there is paucity of information on DCC in small-for-gestational-age (SGA) infants. SGA infants from industrialised countries often manifest an increased incidence of polycythaemia caused by chronic hypoxaemia in utero leading to increased erythropoiesis. These infants are considered to be at greater risk of symptoms and clinical consequences of altered viscosity (Anderson & Hay 1999). However, the baseline risk for polycythaemia-hyperviscosity syndrome in SGA infants from developing countries might be considerably lower. The prevalence of infants born with low cord haemoglobin is high in areas where malaria and iron deficiency anaemia in pregnancy are common (Brabin 1992; le Cessie et al. 2002; Brabin et al. 2004).
Mean bilirubin concentrations were higher in the DCC group, but they did not reach the levels requiring phototherapy or exchange transfusion. We did not observe clinical hyperbilirubinaemia in the first 24 h after birth, although visual estimation of bilirubin levels using the icterometer could lead to misclassification. Neonatal admissions for a longer period to observe later onset of neonatal jaundice was not possible, as early discharge within 24 h after an uncomplicated delivery was expected by mothers and was hospital policy.
We used ZPP levels at birth as a proxy for total body iron contents, and found that normal birth weight infants from Zambia already had low iron stores at birth. ZPP is regarded as a sensitive indicator of iron-deficient erythropoiesis (Rettmer et al. 1999; Juul et al. 2003; Miller et al. 2003; Kling 2006; Lott et al. 2006). Raised ZPP levels indicate incomplete iron incorporation into protoporphyrin, as zinc substitutes for iron when stores are low. Although ZPP has the advantage of low cost and simplicity, its specificity may be low as it can be increased by malaria and other infections (Stoltzfus et al. 2000; Asobayire et al. 2001), which could have caused an overestimation of ZPP values at 6 months, when a small number of infants had positive smears.
Two recently published randomised controlled trials (Ceriani Cernadas et al. 2006; Chaparro et al. 2006) evaluated the effect of cord clamping on maternal blood loss. Major limitations of these trials were diversity in measuring blood loss (visual estimation vs. measuring jar), diversity in mode of delivery (100% vaginal vs. >25% caesarean section) and in definition of DCC. No differences were found between DCC and immediate clamping. In our study there was no significant difference between groups in the midwives’ assessment of maternal bleeding, although the sample size was too small to adequately detect modest differences. A further limitation is that we were not able to quantitatively measure maternal blood loss. However, we tried to quantify the blood loss indirectly by comparing maternal haemoglobin levels in the first stage of labour with levels on average 16 h postpartum and found no differences.
The strengths of this study are its randomised design, the low drop-out rate for a rural area, and the control for confounding factors which could influence infant haematological status, including obstetrical procedures, maternal or infant use of iron supplements, intercurrent (malaria) infections and complementary infant feeding practice. A limitation is that mothers who were already assigned to a treatment group could later become ineligible. This could not be avoided in view of timing and nature of the intervention. In view of this the data are analysed as per protocol and not on an intention-to-treat basis. A second limitation was that the investigators, assessors and mothers were not blinded to the assigned intervention. Theoretically this could have biased the clinical estimation of maternal blood loss, or the clinical evaluation for the hyperviscosity syndrome and hyperbilirubinaemia.
In this trial a sample of cord blood was collected for Hb analysis immediately after clamping and cutting the umbilical cord. This baseline measurement was used to compare the Hb changes between the DCC and control groups. There was a slight baseline imbalance which influenced our chosen summary measure. We corrected for this using the analysis of covariance.
The trial aimed to determine the effect only of DCC on infant anaemia and did not assess other known causes of infant anaemia, including human immunodeficiency virus (HIV) infection. We do not have HIV prevalence estimates for Mpongwe District. HIV prevalence among antenatal clinic attendees 1998–1999 in Kapiri Mposhi District, a neighbouring district, was 8.3% and in Ibenga 9.3% (Fylkesnes et al. 2001). Mpongwe District has a similar low rural population density that is dispersed and village-based.
DCC has previously been shown to be more beneficial in babies of anaemic mothers (Gupta & Ramji 2002; van Rheenen & Brabin 2004). We were not able to confirm this, as maternal anaemia prevalence was lower than expected in this study population. All women had received two intermittent preventive doses of sulphadoxine-pyrimethamine during their pregnancy to control malaria in pregnancy which should reduce maternal anaemia prevalence (Garner & Gulmezoglu 2003). The majority of women in Mpongwe District deliver at home and the prevalence of maternal anaemia could be much higher in these community deliveries. A larger community-based trial with this type of intervention would be difficult to conduct. The relevance of the present study findings for community deliveries needs to be assessed in the context of traditional delivery practices. This is an important area for further research.
Infant anaemia at 6 months was highly prevalent in both the DCC and control groups, indicating that DCC alone was not adequate to prevent this. Combining DCC with intermittent antimalarial treatment in infants and use of impregnated bed nets requires further assessment. Low birth weight and preterm babies have a higher incidence of infant anaemia, because of lower body iron stores and higher iron requirements for catch-up growth. Evaluation of DCC is required in this higher risk group, especially in developing countries, where no studies of DCC in low birth weight babies have been conducted.
We have shown that delaying clamping of the umbilical cord till cord pulsations stop significantly increased Hb levels at the age of 4 months in term babies. The decision to clamp after cessation of pulsations was based on the fact that in many delivery rooms in resource-poor settings a clock is absent or not functioning (own observations). The rate of placental transfusion is markedly influenced by the position of the delivered infant. An infant held 50–60 cm above the placenta will not receive net blood from the placenta (Yao & Lind 1969). From 10 cm above to 10 cm below the level of the placenta, infants receive the maximum possible amount within 3 min of birth. Keeping the infant 40 cm below the placenta hastens placental transfusion to near-completion within 1 min (Yao & Lind 1969; Linderkamp 1982). Cord clamping should be delayed for at least 3 min for the optimal volume of placental transfusion, provided that the position of the term infant before clamping is on the mothers’ abdomen or lower (van Rheenen & Brabin 2006).
Effective interventions are urgently needed to improve child survival in malaria-endemic areas. DCC is a simple, cost-free and safe delivery procedure that might offer a sustainable strategy to reduce early infant anaemia risk when other interventions are not yet feasible. It should be included in integrated programmes aimed at reducing anaemia in young children in developing countries.
We thank the management board of Mpongwe Mission Hospital for their hospitality, maternity and laboratory staff for their assistance; Mr J. Western (Technical Department, Mpongwe Mission Hospital) for maintenance of our equipment; Mr J. de Witte for the logistical support and Prof. Dr H.J. Verkade (University Medical Center Groningen) and Dr F. ter Kuile (Liverpool School of Tropical Medicine) for their critical review of the manuscript. This study received financial support from the Amsterdam–Liverpool Programme for research in Tropical Child Health and from the Bemmel Support Group.