Tactile stimulation in low-risk infants: Results of a systematic Review

Angela Underdown

Associate Professor in Early Childhood, University of Warwick

Jane Barlow

Associate Professor, Reader in Public Health, University of Warwick

Address for Correspondence:

Angela Underdown

WarwickMedicalSchool

University of Warwick

Coventry

CV4 7AL

Tel: 02476 574884

Email:

Key words: tactile stimulation, infants, systematic review

Tactile stimulation in low-risk infants: Results of a systematic Review

Abstract

Touch, an intrinsic part of caring for an infant, powerfully establishes powerful physical and emotional connections, playing and plays an important role in supporting health and well-being, and in particular affect regulation. The objective of this study was to investigate published research to ascertainwhether infant massagetactile stimulationoffers an effective intervention to support infant mental and physical health. ACochrane systematic review into the effects of infant massage on low-risk infants under 6 months identified twenty-two studies (13 from China and 9 from elsewhere) that met our inclusion criteria. Study criteria allowed only for randomised controlled trials to be included. Study quality was variable. Beneficial effects for mother-infant interaction, maternal warmth and reduction of intrusiveness, infant attentiveness, liveliness and happiness were demonstrated in one study; other studies reported no beneficial effect on infant temperament, attachment or development. Individual studies reported beneficial effects on amount of crying, serum levels of norepinephrine and epinephrine, and nocturnal urinary secretion of 6-sulphatoxymelatonine and urinary cortisol. Beneficial effects on patterns and quality of sleep were reported in two studies. Beneficial effects, and on length of sleep were shown in studies from China but not elsewhere in the world. Effects on weight, length and head circumference were evident only in the meta-analysis of studies from China, and in a large study from a Korean orphanage in which normal infant care was highly atypical. It is concluded that there is some evidence that infant massage may have beneficial effects on infants’ sleep and crying patterns, mother-infant interaction and the infant’s physiological response to stress. There is also some evidence that infant massage may be beneficial for infant growth, specifically in situations where infants are deprived of sufficient adult contact, but all of these findings require more robust studies. No evidence was presented that infant massage does harm.

Tactile stimulation in low-risk infants: Results of a systematic Review

Introduction

Touch is an intrinsic part of caring for an infant thatestablishes powerful physical and emotional connectionsbetween carer and baby,and plays a pervasive role in communication and affect regulation(Moszkowski and Stack 2007). For most infants, everyday routines involve many tactile interactions, communicating a range of somatosensory messages including emotional feelings, pressure, temperature, texture, softness or even pain.The amount of touch considered appropriate between adults and infants varies between cultures with some babies experiencing close contact with the mother’s body during most of the first year while others experience extended periods of separation from the mother by being encouraged to sleep alone in the nursery and through the use of day-care (Rogoff 2003).

Maternal sensitivity has been identified as a significant component in the development of an infant’semotional attachment(Ainsworth 1973), and the sensitivity (or otherwise) of a carers’ responses is transmitted to the infant through a combination of touch, voice and facial movements. During the past decade there has been increasing recognition about the specific role of touch or tactile stimulation in the development of the infant’s capacity for affect regulation.

Affect Regulation and the Role of Touch

The formation of psychological attachments appears to be interlinked with chemical changes and functional development in the infant neural system (Schore 2003), and sensitive tactile stimulation during maternal nursing of the infant, is thought to play a significant role in the growth of the dendrites that form the crucial neural connections (Greenough and Black 1992 cited in Schore 2003). In primates, neurobiological research has shown that ‘critical levels of tactile input of specific quality and emotional content is important for normal brain maturation’ (Martin et al 1991:3355). At birth the primary somatosensory centre, which processes tactile and kinesthetic information in the brain’s cerebral hemispheres, is metabolically active (Chugani 1996). The hypothalamic-pituitary-adrenocorticol (HPA) system produces and regulates the glucocorticoid cortisol in response to stress (Gunnar 1998) and for the first two months after birth the infant’s HPA system is highly labile (Gunnar, Brodersen, Krueger and Rigatuso 1996). However, from the age of about two months the infant’s stress systems are becoming organised via transactions with the sensitive main carer who acts as a ‘buffer’ to the reactivity of the HPA system (Nelson and Bosquet 2000).

Touch has been reported as an intrinsic factor in helping to regulate infant behavioural states (Brazelton 1990) and de Chateau (1976) found that infants who had extra bodily contact after birth smiled more and cried less at three months observation. By 8 weeks post-natally the somatosensory connections to the amygdala (the almond shaped group of neurons located deep within the brain’s temporal lobes that have a major role in the processing and memory of emotional reactions) are forming and Schore (2003) suggests that sufficient levels of tactile stimulation releases early pro-attachment behaviour.

It is the security that most children experience within their attachment relationships that is thought to impact upon the high neonatal cortisol responses, so that by the age of one year infants are less likely to show an increase in response to stressors (Gunnar and Donzella 2002). This appears to be due to the development of functional glucocortoid hyporesponsiveness, as children learn to feel secure and safe with caregivers who respond to their cues (Gunnar and Donzella 2002). However, Gunnar and Donzella (2002) further report that elevated cortisol levels have been noted when children are exposed to moderately less sensitive and responsive care. These findings are supported by ethological studies (Liu et al 1997) which have demonstrated that the amount of tactile stimulation (maternal licking and grooming) experienced by rodent pups

modulates cortisol production.Harlow’s famous studies alsodemonstrated the importance of tactile experience by replacing the rhesus monkey mothers with wire mesh dummies, one dispensing milk from a bottle and the other clad in terry toweling cloth (Harlow 1958). The monkeys clung to the tactile surrogate mother when fearful and were more confident to explore when it was present (Harlow 1958). In other studies Harlow (1965; 1971) demonstrated that isolated monkeys developed abnormal social behaviour and later failed to care for their young.

Other studies (Mitchell 1970; Goosen 1981; Kraemer 1985) have producedsimilar findings of abnormal social and motor behaviour in socially deprived monkeys. More recent work (Martin et al 1991) investigating the neurobiological mechanisms underlying the behaviour of socially deprived monkeys found that social/sensory deprivation of non-human primates in the first year of life have pronounced alterations in the organization and patterned arrangements of the basal ganglia neurotransmitters. The basal ganglia are a group of subcortical nuclei involved in motor control, cognition, and emotion (Carver and Carver 2003) and abnormalities in this region are characterised by abnormal movements and a number of neuropsychiatric disorders (Carver and Carver 2003). Evidence from neurobiological research on primates therefore suggests that “in early post-natal life, maintenance of critical levels of tactile input of specific quality and emotional content is important for normal brain maturation” (Martin 1991:3355).

Research with human infants has been restricted to naturallyoccurring situations where children have been raised in highly adverse conditions. Children living in Romanian orphanages, who lacked social and physical stimulation and opportunities to form emotional attachments with consistent care-givers had complex ‘flattened’ cortisol profiles that did not show the expected daily rhythm, with raised levels in the morning (Carlson et al 1995; Carlson et Earls 1997).This finding suggests that HPA axis activity may have been adversely affected (Gunnar and Vazquez 2001), and although no firm conclusions were possible from this data (Gunnar and Donzella 2002), other evidence shows that children raised in orphanages of this type have stunted growth (Johnson 2000). Elevated corticotrophin-releasing hormones (CRH) are thought to be instrumental in this type of faltering growth (Johnson et al 1992).

Intervening using tactile stimulation

Tactile stimulation or infant massage is widely practiced in many parts of the world, especially Africa, Asia and the South Pacific (Field, Schanberg, Davalos, Malphurs 1996). A survey of 332 primary caretakers of neonates in Bangladesh, for example, found that 96% engaged in massage of the infant's whole body between one and three times daily (Darmstadt et al 2002).

In the West, tactile stimulation has until recently been largely restricted to high-risk infantsoften in intensive care settings, in which the amount of handling that an infant received was thought to be suboptimal (Field 2000 cited in Zeanah 2000; Vickers 2000). Increasingly, however, parents in western cultures are being taught the techniques of tactile stimulation or infant massage (Underdown cited in Barlow and Svanberg in press) often by attending a weekly class. Claims for the effects of such stimulation are wide and include benefitsfor parent-infant interaction, sleep, respiration, elimination and the reduction of colic and wind(Blackwell 2000; Field 2000 cited in Zeanah 2000). The aim of this review was to address whether tactile stimulation in medically low-risk infants could be shown to have an impact on either the mental well-being of infants or the relationship between the mother and the baby. Data on infant growth has also been included, due to the complexity of the relationship between growth and other aspects of mental health.

Method

A systematic review of randomised controlled trials was conducted. Searches were undertaken of a range of electronic databases includingCENTRAL 2005 (Issue 3), MEDLINE (1970 to 2005), PsycINFO (1970 to 2005), CINAHL (1982 to 2005), EMBASE (1980 to 2005), and a number of other Western and Chinese

databases. We only included studies that had examined the impact of any form of tactile stimulation on low-risk infants under the age of six months. The intervention was defined as ‘systematic tactile stimulation by human hands’ (Vickers 2004) and low-risk infants were defined as having no apparent physical health adversities. The selection criteria were established prior to the systematic review searches, so that clarity would be ensured about which studies should be included. The actual selection of studies caused much debate and conferring between the reviewers as ‘low risk’ in infancy is open to definition and it is usual to consider risk within relationships rather than purely physical challenges to the infant only. These strict selection criteriaStudies comprising led to infants with infants ofpost-natally depressed mothers (Field et al 1996; Onozawa et al 2001) and those being cared for in an orphanage (Kim et 2003)being were included (i.e. as they were reported as beingon healthy full-term infants, whereas infants who were HIV exposed (Scafidi et al 1996) were not includedexcluded. Studies were included if infants were randomised to either a tactile stimulation group or a control group that receivedno intervention or standard care. No language restrictions were applied.

Data management

Data were extracted, checked and entered into RevMan version 4.2 (Cochrane Collaboration, 2003) by two reviewers independently. Discrepancies were discussed with a third reviewer.

Assessment of quality and risk of bias

Trials were assigned a quality category based on allocation concealment. Aspects of study design that may increase bias were also appraised (sample size, number of infants lost to follow-up, the method of dealing with attrition/drop out, use of blinding to assess outcomes, and whether there was any assessment of the distribution of confounders).

Data synthesis

Using a random effect model we report the differences in continuous outcomes between the treatment and control groups as weighted mean differences with 95% confidence interval. In the case of measures where data were reported on incompatible scales, we present the standardised mean difference and 95% confidence intervals. Where it was not possible to synthesise the data, effect sizes and 95% confidence intervals have been calculated for individual outcomes in each study. In the absence of data with which to calculate effect sizes we report the significance level presented in the primary study. A minus sign indicates a result favouring the intervention group.

Due to concerns about uniformly positive results, inadequate information about the design and conduct, and the absence of any reported dropout, a post hoc decision was taken to analyse separately the results of thirteen studies undertaken in China (Liu Chun Li 2005; Liu 2005; Lu Jiao, Li Ju Zhan, Wu Li Fang 2005; Na Zhuo Hua, Xie Hui Yun, Huang Jian Hua 2005; Shao L et al.2005; Sun Hai Yun, Gao Xiang Yu, Zhao Xue Mei 2005; Ye Hong Yun 2004; Xua Li Shuan, Qing Gui Romg, Ye Mei Yan et al 2004; Duan Lihong, Li Weihong, Shi Fentao 2002; Shi Li, Xue, Rong 2002; Ke, Ling, Li. 2001; Zhai, Pan Xian, Hua et al 2001; Wang Bin, Shen Yue Hua, Jin Run Yan et al 1999).

Sensitivity Analysis

A sensitivity analysis was undertaken on the non-Chinese studies to assess the impact on the findings of one large Korean study (Kim, Shin, White-Traut 2003) of infants receiving orphanage care.

Results

We identified 809 abstracts in Western databases, of which 35 were obtained for review. Ten studies met the inclusion criteria, one (Koniak Griffin, Ludington-Hoe et al 2001) of which reported follow-up of one of the other nine and one of which was carried out in China (Duan Lihong, Li Weihong, Shi Fentao 2002). A hand-

search of references resulted in the identification of one further study (Ke, Ling, Li 2001).

Of the abstracts reviewed from the Chinese databases,12 studies were identified as suitable for inclusion, producing in total 13 Chinese studies (Liu Chun Li 2005; Liu 2005; Lu Jiao, Li Ju Zhan, Wu Li Fang 2005; Na Zhuo Hua, Xie Hui Yun, Huang Jian Hua 2005; Shao L et al.2005; Sun Hai Yun, Gao Xiang Yu, Zhao Xue Mei 2005; Ye Hong Yun 2004; Xua Li Shuan, Qing Gui Romg, Ye Mei Yan et al 2004; Duan Lihong, Li Weihong, Shi Fentao 2002; Shi Li, Xue, Rong 2002; Ke, Ling, Li. 2001; Zhai, Pan Xian, Hua et al 2001; Wang Bin, Shen Yue Hua, Jin Run Yan et al 1999)

In addition to the Chinese studies,one Korean (Kim, Shin, White-Traut 2003), one Israeli (Goldstein Ferber et al 2002),one British (Onozawa, Glover et al 2001) one Indian (Argawal, Ashish et al 2000), and five North American (Elliot, Reilly 2002; Cigales, Field et al 1996; Jump 1998; Field, Grizzle et al 1996; Koniak Griffin, Ludington-Hoe 1988)studies were included. One follow-up study was also includedand one follow-up (Koniak Griffin, Ludington-Hoe et al 1995) were included.

Table one shows the characteristics of the included studies.

Table 1

Types of studies

A range of different tactile stimulation or massage methods were reviewed. In four studies (Argawal 2000; Elliott 2002; GoldsteinFerber 2002;Koniak-Griffin 1988) parents were taught massage techniques prior to them conducting massage on their infants in the home. In other two studies massage was offered by research associates (Cigales 1997;Field 1996). In the Kim (2003) study, orphans received a multi-modal intervention of massage, talking and eye contact from research associates who were trained to be responsive to the infants’ responses. Although it was not possible to isolate the effects of eye contact and talking, thisstudy was included because both these components are an intrinsic part of some included infant massage programmes. In the 13 Chinese studies the massage was mostly administered by a nurse with specialist training in infant massage, following which the technique was taught to the parents who continued the massage at home.

Types of massage
The massage programmes evaluated in the included studies varied in terms of duration and frequency. In one study, infants received a daily 30 minutes intervention over 14 days (GoldsteinFerber 2002). In the Kim (2003) study infants were massaged 15 minutes, twice daily for 4 weeks. In further studies, infants received 10 minutes of massage daily over a four week period (Argawal 2000) or a minimum of 10 minutes massage daily over 16 weeks (Elliott 2002). In the Field (1996) study infants received 15 minutes of massage twice weekly over a period of six weeks and in the Koniak-Griffin (1988) study infants received 5-7 minutes of massage daily over 3 months. In two studies (Jump 1998 and Onozawa 2001) mothers were taught massage techniques for approximately an hour per week as part of four, weekly, group-based sessions, following which the continuation of this practice at home varied according to parental motivation. In the 13 Chinese studies infants were mostly massaged for fifteen minute periods up to three times a day over a period extending up to 6 weeks (Liu 2005; Liu Chun 2005; Lu 2005; Na 2005; Shao 2005; Sun 2004; Ye 2004; Xua 2004; Duan 2002; Shi 2002; Ke 2001; Zhai 2001; Wang 1999). In the Cigales study (1997) massage was administered only once prior to the conducting of an experimental task to assess the impact of massage on cognitions.

Findings

Stress hormones

One USA US study (Field et al 1996) showed reduced levels of norepinephrine (weighted mean difference -60.30ng/gm creatinine 95% confidence intervals -111.88 to -8.72) and epinephrine (weighted mean difference -13.00 ng/gm creatinine 95% confidence intervals -20.08 to -5.29) among massaged infants, but no effect on levels of serotonin (weighted mean difference -295.50ng/gm creatinine 95% confidence interval -705.25 to 114.25). This study also measured salivary cortisol levels 20-minutes post-intervention and showed no difference between groups (weighted mean difference -0.20 ng/dl 95% confidence interval

-4.60 to 4.20) (Field et al 1996). However, urinary cortisol levels measured later using radioimmune assay, were lower in the massage group (weighted mean difference -360.40 ng/gm creatinine 95% confidence intervals -633.79 to -87.01).

Circadian Rhythms