Downloaded from http://fn.bmj.com/ on June 6, 2015 - Published by group.bmj.com
Original article
Effect of emollient therapy on clinical outcomes in preterm neonates in Pakistan: a randomised controlled trial Rehana A Salam,1 Gary L Darmstadt,2 Zulfiqar A Bhutta3,4 1
Division of Women & Child Health, The Aga Khan University, Karachi, Sindh, Pakistan 2 Global Development Division, Bill & Melinda Gates Foundation, Seattle, Washington, USA 3 Center of Excellence in Women & Child Health, The Aga Khan University, Karachi, Pakistan 4 Centre for Global Child Health, The Hospital for Sick Children, Toronto, Canada Correspondence to Professor Zulfiqar A Bhutta, Center of Excellence in Women & Child Health, The Aga Khan University, Karachi 74800, Pakistan, and Centre for Global Child Health, The Hospital for Sick Children, Toronto, Canada; zulfi[email protected] Received 12 July 2014 Revised 1 December 2014 Accepted 31 December 2014 Published Online First 30 January 2015
To cite: Salam RA, Darmstadt GL, Bhutta ZA. Arch Dis Child Fetal Neonatal Ed 2015;100: F210–F215. F210
ABSTRACT Importance Newborn oil massage, a traditional community practice, could potentially benefit thermoregulation and skin barrier function, and prevent serious infections, morbidity and mortality in high-risk preterm infants, but has only been evaluated in limited studies in low income settings. Objectives To assess the efficacy of topical coconut oil applications among a cohort of hospital-born preterm infants. Design A prospective, individually randomised controlled clinical trial. Setting Nursery and neonatal intensive care unit at Aga Khan University Hospital, Pakistan. Participants Of 270 eligible neonates, a consecutive cohort of 258 hospital-born preterm infants (gestational age ≥26 weeks and ≤37 weeks). Intervention Twice daily topical application of coconut oil by nurses from birth until discharge and continued thereafter by mothers at home until completion of the 28th day of life. Primary outcome measures Incidence of hospitalacquired bloodstream infections. Secondary outcome measures Weight gain, skin condition and neonatal mortality. Results 23% of the enrolled neonates developed clinically suspected sepsis while 14% developed blood culture proven infection. The unadjusted hazard for developing hospital-acquired infection in the control group was 4.7 (95% CI 1.8 to 12.4) compared with the intervention group. After adjusting for gestational age, birth weight, duration of intubation and duration of hospitalisation for possible confounding, the hazard for hospital-acquired infection in the control group was 6.0 (95% CI 2.3 to 16) compared with the intervention group. The rate of hospital-acquired infections in the control and intervention groups was 219.1 and 39.5 per 1000 patient-days, respectively. Mean weight gain was 11.3 g/day higher (95% CI 8.1 to 14.6, p26 and 750 g were potentially eligible for the study. Any preterm infant admitted with severe respiratory distress syndrome (as certified by the consulting physician on the basis of clinical and radiological findings21), life-threatening congenital abnormality, hydrops fetalis, congenital skin anomalies or signs of skin infection ( pustulosis) or for a major surgical procedure was excluded. Although the NICU staff ( physicians and nurses) could not be blinded to intervention group allocation, they were unaware of the randomisation process. Baseline blood cultures were sent for all enrolled infants, and baseline information was collected on maternal characteristics, pregnancy history, mode of delivery, and immediate postnatal care practices
and events. Weight was recorded and gestational age estimated from ultrasound records and Ballard scores at admission. On admission to the nursery or NICU, the lead investigator (RAS) established eligibility with the preset criteria for every newborn infant, obtained written informed consent from the parent or guardian, and randomly allocated the newborns to either the CO group or the control group by sequential allocation using a block randomisation list. The study physicians were blinded to the randomisation procedure, but not the nature of the intervention. If the blood cultures obtained on admission came back positive for any pathogenic organism, the infant was censored from analysis. CO was procured locally from a single source supplier and aliquots were prepared in the NICU for dispensing in small (30 mL) individual sterile containers twice weekly to ensure compliance by the principal investigator. CO was applied by the assigned nurse to the newborns in the intervention group. Nurses received 3 days training on study procedures, including how to apply CO in a standardised manner that minimised skin trauma or contamination. Following hand-washing, CO (approximately 5 mL/kg) was gently applied twice a day over the newborn’s body surface (except the face and scalp and intravenous infusion sites) over 2–3 min. CO application was coordinated along with routine tasks of the nurses in order to prevent excessive handling of the baby and treatment was provided within the incubators to prevent hypothermia. Newborns in the control group received routine skin care. Routine infection control precautions and hand-washing before and after contact with the newborn were emphasised throughout the study period as per hospital policy.
Sample size and data analysis The estimated rate of blood culture proven hospital-acquired infection among all preterm infants admitted to AKUH in 2010 was 28% (Bhutta ZA, unpublished data). We estimated that we would require 130 preterm infants in each arm based on a hypothesised 40% reduction in the incidence of hospital-acquired infections22 to achieve 80% power and 5% significance.23
Clinical care and data collection for morbidity and mortality monitoring Other than emollient use, all newborns in the control and intervention groups received the same routine care. Indwelling intravascular catheters, oxygen therapy and tube feeding were continued as ordered by the physicians in both the groups. Newborns on ventilators received routine ventilator care according to the unit protocol. All newborns in the NICU were monitored continuously for heart rate, oxygen saturation and blood pressure. Vital signs and fluid intake/output charts were maintained every hour for intubated newborns and two hourly for non-intubated newborns. Decisions regarding treatment with antibiotics and their duration for sick newborns were made by the attending physicians and nurses as per standard AKUH policy.24 25 Blood cultures were sent whenever clinical signs of suspected systemic infection were apparent, including lethargy, poor peripheral perfusion, fever or temperature instability, increased demand for oxygen, abdominal distention, feeding intolerance, bradycardia or apnoea.
Ambulatory follow-up Newborn infants in both study groups were followed weekly after hospital discharge until the 28th completed day after birth. Mothers of newborn infants in the intervention group were
Salam RA, et al. Arch Dis Child Fetal Neonatal Ed 2015;100:F210–F215. doi:10.1136/archdischild-2014-307157
F211
Downloaded from http://fn.bmj.com/ on June 6, 2015 - Published by group.bmj.com
Original article trained to apply CO twice daily at home until the 28th completed day of age and were provided a weekly supply of CO in a sterile container during follow-up visits. Compliance with CO application was assessed at follow-up visits by determining the volume of oil required to replenish the previous container. At each visit, the infant’s weight was recorded, and skin condition and clinical status were documented.
Outcome measures The primary outcome was incidence of hospital-acquired bloodstream infection, defined as having a blood culture positive for any pathogenic organism. Blood cultures were sent whenever the physician suspected systemic infection based on clinical signs, as defined above. Secondary outcomes evaluated included weight gain and skin condition, which were assessed each day after the sponge bath in both groups by nurses. Skin condition of the neonates was compared according to the following grading criteria: 0: Intact; 1: Redness; 2: Visible skin flaking and prick marks; and 3: Bruises and abrasions. Mortality at the 28th day of age was also recorded.
Statistical methods Analysis was done on an intention-to-treat basis. All analyses were done using SPSS V.11.5 (SPSS Inc, Chicago, Illinois, USA). Data were collected for baseline group characteristics and potential confounding variables including gestational age, birth weight, duration of intubation and duration of hospitalisation. Adjusted relative risk of hospital-acquired infection and mortality was calculated with 95% CIs using Cox regression. Survival analysis was conducted for hospital-acquired infection and mortality and reported as HRs and 95% CIs. The mean daily weight gain in both the groups was compared using a t test while neonates with altered skin integrity were compared separately for each grade of disruption (according to the criteria described) using the χ2 test of association.
RESULTS A total of 3361 deliveries occurred at AKUH during the study period, of which 270 were preterm (8.0%) and were screened for eligibility (figure 1). Parents of seven newborns refused to provide written consent, three had severe respiratory distress syndrome and two had fatal congenital heart anomalies and were excluded. The remaining 258 newborns were randomised to either the intervention (n=128) or control groups (n=130). Two infants in each group were subsequently found to have positive blood cultures for a pathogenic organism at the time of admission and were censored from the intention-to-treat analysis. Baseline characteristics of the newborns in both the groups were comparable in terms of maternal and infant characteristics except for birth weight where the control group infants were slightly larger ( p=0.05) (table 1). Overall, 23% (60 of 258) of the enrolled neonates were clinically suspected to have sepsis and their blood cultures were sent, while 14% (36 of 258) developed blood culture proven infection during hospitalisation (table 2). All the neonates had a single episode of confirmed infection. Out of blood culture proven infections, significantly fewer infants developed infection in the intervention group (7%, nine of 128) compared with the control group (21%, 27 of 130). The rate of positivity of blood cultures was also lower in the intervention (40.9%, nine of 22) compared with the control (71.1%, 27 of 38). The unadjusted hazard of developing hospital-acquired infection in the control group was 4.7 (95% CI 1.8 to 12.4). After adjusting for gestational age, birth weight, duration of intubation and duration of hospitalisation for possible confounding, the hazard of hospital-acquired infection was 6.0 (2.3–16) (figure 2). Incidence of hospital-acquired infection in the intervention and control groups was 219.1 and 39.5 per 1000 patient days, respectively. More than half (56% (20 of 36)) of the hospital-acquired infections were caused by gram-negative bacilli that included Acinetobacter spp, Klebseilla pneumoniae, Pseudomonas spp and Enterococcus spp; 14% (five of 36) were
Figure 1 Trial flow chart.
F212
Salam RA, et al. Arch Dis Child Fetal Neonatal Ed 2015;100:F210–F215. doi:10.1136/archdischild-2014-307157
Downloaded from http://fn.bmj.com/ on June 6, 2015 - Published by group.bmj.com
Original article Table 1 Baseline characteristics of enrolled infants Variables Maternal Primigravida Parity Mode of delivery Spontaneous vaginal delivery Caesarean section Eclampsia Gestational diabetes Antenatal corticosteroids Infant Age at enrolment (mean hours of life (SD)) Sex Male Female Gestational age (weeks) Birth weight (grams) Neonatal intensive care unit admission Severity of illness Ventilated on admission Respiratory distress syndrome Meconium aspiration syndrome Intrauterine growth retardation Transient tachypnoea of neonates
Intervention n=128
Control n=130
49 (38.9%) 1.2 (1.443)
51 (39.8%) 1.1 (1.35)
35 (27.8%) 91 (72.2%) 14 (11.1%) 7 (5.6%) 90 (71.4)
43 85 12 10 97
25.5 (6.8) 57 71 33.9 2114.1 61
(33.6%) (66.4%) (9.4%) (7.8%) (75.8%)
25.0 (4.9)
(44.4%) (55.6%) (2.4) (499.7) (48.4)
60 70 34.4 2249.5 50
(46.1%) (53.9%) (2.6) (569.0) (39.1%)
28 (21.9%) 48 (37.5%) 02 (1.6%) 03(2.3%) 04 (3.1%)
23 42 02 02 03
(17.7%) (32.3%) (1.5%) (1.5%) (2.3%)
gram-positive organisms including Staphylococcus aureus and Streptococcus spp; 8% (three of 36) were Candida albicans; and 22% (eight of 36) were coagulase-negative Staphylococcus spp (table 2). A significantly greater proportion of newborn infants were switched from first to second-line antibiotics in the control group compared with the CO application group ( p
Original article
Effect of emollient therapy on clinical outcomes in preterm neonates in Pakistan: a randomised controlled trial Rehana A Salam,1 Gary L Darmstadt,2 Zulfiqar A Bhutta3,4 1
Division of Women & Child Health, The Aga Khan University, Karachi, Sindh, Pakistan 2 Global Development Division, Bill & Melinda Gates Foundation, Seattle, Washington, USA 3 Center of Excellence in Women & Child Health, The Aga Khan University, Karachi, Pakistan 4 Centre for Global Child Health, The Hospital for Sick Children, Toronto, Canada Correspondence to Professor Zulfiqar A Bhutta, Center of Excellence in Women & Child Health, The Aga Khan University, Karachi 74800, Pakistan, and Centre for Global Child Health, The Hospital for Sick Children, Toronto, Canada; zulfi[email protected] Received 12 July 2014 Revised 1 December 2014 Accepted 31 December 2014 Published Online First 30 January 2015
To cite: Salam RA, Darmstadt GL, Bhutta ZA. Arch Dis Child Fetal Neonatal Ed 2015;100: F210–F215. F210
ABSTRACT Importance Newborn oil massage, a traditional community practice, could potentially benefit thermoregulation and skin barrier function, and prevent serious infections, morbidity and mortality in high-risk preterm infants, but has only been evaluated in limited studies in low income settings. Objectives To assess the efficacy of topical coconut oil applications among a cohort of hospital-born preterm infants. Design A prospective, individually randomised controlled clinical trial. Setting Nursery and neonatal intensive care unit at Aga Khan University Hospital, Pakistan. Participants Of 270 eligible neonates, a consecutive cohort of 258 hospital-born preterm infants (gestational age ≥26 weeks and ≤37 weeks). Intervention Twice daily topical application of coconut oil by nurses from birth until discharge and continued thereafter by mothers at home until completion of the 28th day of life. Primary outcome measures Incidence of hospitalacquired bloodstream infections. Secondary outcome measures Weight gain, skin condition and neonatal mortality. Results 23% of the enrolled neonates developed clinically suspected sepsis while 14% developed blood culture proven infection. The unadjusted hazard for developing hospital-acquired infection in the control group was 4.7 (95% CI 1.8 to 12.4) compared with the intervention group. After adjusting for gestational age, birth weight, duration of intubation and duration of hospitalisation for possible confounding, the hazard for hospital-acquired infection in the control group was 6.0 (95% CI 2.3 to 16) compared with the intervention group. The rate of hospital-acquired infections in the control and intervention groups was 219.1 and 39.5 per 1000 patient-days, respectively. Mean weight gain was 11.3 g/day higher (95% CI 8.1 to 14.6, p26 and 750 g were potentially eligible for the study. Any preterm infant admitted with severe respiratory distress syndrome (as certified by the consulting physician on the basis of clinical and radiological findings21), life-threatening congenital abnormality, hydrops fetalis, congenital skin anomalies or signs of skin infection ( pustulosis) or for a major surgical procedure was excluded. Although the NICU staff ( physicians and nurses) could not be blinded to intervention group allocation, they were unaware of the randomisation process. Baseline blood cultures were sent for all enrolled infants, and baseline information was collected on maternal characteristics, pregnancy history, mode of delivery, and immediate postnatal care practices
and events. Weight was recorded and gestational age estimated from ultrasound records and Ballard scores at admission. On admission to the nursery or NICU, the lead investigator (RAS) established eligibility with the preset criteria for every newborn infant, obtained written informed consent from the parent or guardian, and randomly allocated the newborns to either the CO group or the control group by sequential allocation using a block randomisation list. The study physicians were blinded to the randomisation procedure, but not the nature of the intervention. If the blood cultures obtained on admission came back positive for any pathogenic organism, the infant was censored from analysis. CO was procured locally from a single source supplier and aliquots were prepared in the NICU for dispensing in small (30 mL) individual sterile containers twice weekly to ensure compliance by the principal investigator. CO was applied by the assigned nurse to the newborns in the intervention group. Nurses received 3 days training on study procedures, including how to apply CO in a standardised manner that minimised skin trauma or contamination. Following hand-washing, CO (approximately 5 mL/kg) was gently applied twice a day over the newborn’s body surface (except the face and scalp and intravenous infusion sites) over 2–3 min. CO application was coordinated along with routine tasks of the nurses in order to prevent excessive handling of the baby and treatment was provided within the incubators to prevent hypothermia. Newborns in the control group received routine skin care. Routine infection control precautions and hand-washing before and after contact with the newborn were emphasised throughout the study period as per hospital policy.
Sample size and data analysis The estimated rate of blood culture proven hospital-acquired infection among all preterm infants admitted to AKUH in 2010 was 28% (Bhutta ZA, unpublished data). We estimated that we would require 130 preterm infants in each arm based on a hypothesised 40% reduction in the incidence of hospital-acquired infections22 to achieve 80% power and 5% significance.23
Clinical care and data collection for morbidity and mortality monitoring Other than emollient use, all newborns in the control and intervention groups received the same routine care. Indwelling intravascular catheters, oxygen therapy and tube feeding were continued as ordered by the physicians in both the groups. Newborns on ventilators received routine ventilator care according to the unit protocol. All newborns in the NICU were monitored continuously for heart rate, oxygen saturation and blood pressure. Vital signs and fluid intake/output charts were maintained every hour for intubated newborns and two hourly for non-intubated newborns. Decisions regarding treatment with antibiotics and their duration for sick newborns were made by the attending physicians and nurses as per standard AKUH policy.24 25 Blood cultures were sent whenever clinical signs of suspected systemic infection were apparent, including lethargy, poor peripheral perfusion, fever or temperature instability, increased demand for oxygen, abdominal distention, feeding intolerance, bradycardia or apnoea.
Ambulatory follow-up Newborn infants in both study groups were followed weekly after hospital discharge until the 28th completed day after birth. Mothers of newborn infants in the intervention group were
Salam RA, et al. Arch Dis Child Fetal Neonatal Ed 2015;100:F210–F215. doi:10.1136/archdischild-2014-307157
F211
Downloaded from http://fn.bmj.com/ on June 6, 2015 - Published by group.bmj.com
Original article trained to apply CO twice daily at home until the 28th completed day of age and were provided a weekly supply of CO in a sterile container during follow-up visits. Compliance with CO application was assessed at follow-up visits by determining the volume of oil required to replenish the previous container. At each visit, the infant’s weight was recorded, and skin condition and clinical status were documented.
Outcome measures The primary outcome was incidence of hospital-acquired bloodstream infection, defined as having a blood culture positive for any pathogenic organism. Blood cultures were sent whenever the physician suspected systemic infection based on clinical signs, as defined above. Secondary outcomes evaluated included weight gain and skin condition, which were assessed each day after the sponge bath in both groups by nurses. Skin condition of the neonates was compared according to the following grading criteria: 0: Intact; 1: Redness; 2: Visible skin flaking and prick marks; and 3: Bruises and abrasions. Mortality at the 28th day of age was also recorded.
Statistical methods Analysis was done on an intention-to-treat basis. All analyses were done using SPSS V.11.5 (SPSS Inc, Chicago, Illinois, USA). Data were collected for baseline group characteristics and potential confounding variables including gestational age, birth weight, duration of intubation and duration of hospitalisation. Adjusted relative risk of hospital-acquired infection and mortality was calculated with 95% CIs using Cox regression. Survival analysis was conducted for hospital-acquired infection and mortality and reported as HRs and 95% CIs. The mean daily weight gain in both the groups was compared using a t test while neonates with altered skin integrity were compared separately for each grade of disruption (according to the criteria described) using the χ2 test of association.
RESULTS A total of 3361 deliveries occurred at AKUH during the study period, of which 270 were preterm (8.0%) and were screened for eligibility (figure 1). Parents of seven newborns refused to provide written consent, three had severe respiratory distress syndrome and two had fatal congenital heart anomalies and were excluded. The remaining 258 newborns were randomised to either the intervention (n=128) or control groups (n=130). Two infants in each group were subsequently found to have positive blood cultures for a pathogenic organism at the time of admission and were censored from the intention-to-treat analysis. Baseline characteristics of the newborns in both the groups were comparable in terms of maternal and infant characteristics except for birth weight where the control group infants were slightly larger ( p=0.05) (table 1). Overall, 23% (60 of 258) of the enrolled neonates were clinically suspected to have sepsis and their blood cultures were sent, while 14% (36 of 258) developed blood culture proven infection during hospitalisation (table 2). All the neonates had a single episode of confirmed infection. Out of blood culture proven infections, significantly fewer infants developed infection in the intervention group (7%, nine of 128) compared with the control group (21%, 27 of 130). The rate of positivity of blood cultures was also lower in the intervention (40.9%, nine of 22) compared with the control (71.1%, 27 of 38). The unadjusted hazard of developing hospital-acquired infection in the control group was 4.7 (95% CI 1.8 to 12.4). After adjusting for gestational age, birth weight, duration of intubation and duration of hospitalisation for possible confounding, the hazard of hospital-acquired infection was 6.0 (2.3–16) (figure 2). Incidence of hospital-acquired infection in the intervention and control groups was 219.1 and 39.5 per 1000 patient days, respectively. More than half (56% (20 of 36)) of the hospital-acquired infections were caused by gram-negative bacilli that included Acinetobacter spp, Klebseilla pneumoniae, Pseudomonas spp and Enterococcus spp; 14% (five of 36) were
Figure 1 Trial flow chart.
F212
Salam RA, et al. Arch Dis Child Fetal Neonatal Ed 2015;100:F210–F215. doi:10.1136/archdischild-2014-307157
Downloaded from http://fn.bmj.com/ on June 6, 2015 - Published by group.bmj.com
Original article Table 1 Baseline characteristics of enrolled infants Variables Maternal Primigravida Parity Mode of delivery Spontaneous vaginal delivery Caesarean section Eclampsia Gestational diabetes Antenatal corticosteroids Infant Age at enrolment (mean hours of life (SD)) Sex Male Female Gestational age (weeks) Birth weight (grams) Neonatal intensive care unit admission Severity of illness Ventilated on admission Respiratory distress syndrome Meconium aspiration syndrome Intrauterine growth retardation Transient tachypnoea of neonates
Intervention n=128
Control n=130
49 (38.9%) 1.2 (1.443)
51 (39.8%) 1.1 (1.35)
35 (27.8%) 91 (72.2%) 14 (11.1%) 7 (5.6%) 90 (71.4)
43 85 12 10 97
25.5 (6.8) 57 71 33.9 2114.1 61
(33.6%) (66.4%) (9.4%) (7.8%) (75.8%)
25.0 (4.9)
(44.4%) (55.6%) (2.4) (499.7) (48.4)
60 70 34.4 2249.5 50
(46.1%) (53.9%) (2.6) (569.0) (39.1%)
28 (21.9%) 48 (37.5%) 02 (1.6%) 03(2.3%) 04 (3.1%)
23 42 02 02 03
(17.7%) (32.3%) (1.5%) (1.5%) (2.3%)
gram-positive organisms including Staphylococcus aureus and Streptococcus spp; 8% (three of 36) were Candida albicans; and 22% (eight of 36) were coagulase-negative Staphylococcus spp (table 2). A significantly greater proportion of newborn infants were switched from first to second-line antibiotics in the control group compared with the CO application group ( p
