THE JOURNAL OF

CHILD PSYCHOLOGY AND PSYCHIATRY

doi:10.1111/j.1469-7610.2007.01833.x

Journal of Child Psychology and Psychiatry 49:2 (2008), pp 115123

The effect of cleft lip and palate, and the timing of lip repair on motherinfant interactions and infant development
Lynne Murray,1 Francoise Hentges,1 Jonathan Hill,2 Janne Karpf,1 Beejal Mistry,1 Marianne Kreutz,2 Peter Woodall,2 Tony Moss,3 Tim Goodacre,4 and the Cleft Lip and Palate Study team5
1

University of Reading, UK; 2University of Liverpool, UK; 3St Georges Hospital, London, UK; 4John Radcliffe Hospital, Oxford, UK; 5Other members of the Cleft Lip and Palate Study Team: Barry Lovius, Sanu Desai, Brian Christie, Richard Bryson, Peter Budny, Joyce Russell and Roger Green

Background: Children with cleft lip and palate are at risk for psychological problems. Difculties in motherchild interactions may be relevant, and could be affected by the timing of lip repair. Method: We assessed cognitive development, behaviour problems, and attachment in 94 infants with cleft lip (with and without cleft palate) and 96 non-affected control infants at 18 months; motherinfant interactions were assessed at two, six and 12 months. Index infants received either early, neonatal, lip repair, or late repair (34 months). Results: Index infants did not differ from controls on measures of behaviour problems or attachment, regardless of timing of lip repair; however, infants having late lip repair performed worse on the Bayley Scales of Mental Development; the cognitive development of early repair infants was not impaired. Difculties in early motherinfant interactions mediated the effects of late lip repair on infant cognitive outcome. Conclusions: Early interaction difculties between mothers and infants having late repair of cleft lip are associated with poor cognitive functioning at 18 months. Interventions to facilitate motherinfant interactions prior to surgical lip repair should be explored. Keywords: Cleft lip and palate, motherinfant interactions, cognitive development, attachment, behaviour problems, infancy, intelligence, sensitive period. Abbreviations: MDI: Mental Development Index.

Psychological development of children with cleft lip and palate

Cleft lip and palate affects one in every 650 infants (Derijcke, Eerens, & Carels, 1996). Studies of affected children have often been subject to methodological problems (e.g., small, heterogeneous samples, lack of normal controls); nevertheless, ndings concerning psychological effects of clefts have been broadly consistent. In the school years, affected children are at raised risk for both socioemotional and cognitive difculties (Millard & Richman, 2001). In infancy, socio-emotional problems, as measured by infantmother attachment, are not so evident (Wasserman, Lennon, Allen, & Shilansky, 1987; Koomen & Hoeksma, 1993; Speltz, Endriga, Fisher, & Mason, 1997; Maris, Endriga, Speltz, Jones, & DeKlyen, 2000), although cognitive decits do appear to be a particular risk (Starr, Chinsky, Canter, & Meier, 1977; Jocelyn, Penko, & Rode, 1996; Kapp-Simon & Krueckeberg, 2000; Speltz et al., 2000).

de Mello, & Martines, 2006). Giving birth to an infant with a cleft is often followed by shock and distress (Bradbury & Hewison, 1994), and this may affect parentinfant interactions. Previous studies of motherinfant interactions in cleft samples have generally been conducted before lip repair. While caution is required in considering previous ndings, as raters were not blind to cleft status, impairments in maternal responsiveness have been reported (Field & Vega-Lahr, 1984; Wasserman et al., 1987; Speltz, Goodell, Endriga, & Clarren, 1994). Notably, both Wasserman and Allen (1985) and Speltz and colleagues (2000) found that observed maternal interaction difculties predicted lower IQ in affected children.

The timing of lip repair

One factor that may affect parentinfant interactions in the context of infant cleft lip is the timing of surgical repair, since this effects dramatic cosmetic change (Slade, Emerson, & Freedlander, 1999). Although repair is generally performed around 34 months postpartum in the UK, in the past two decades some centres have conducted neonatal repair. Surgical outcome appears unaffected by this difference in timing (Goodacre, Hentges, Moss, Short, & Murray, 2004); however, the psychological consequences of early vs. later lip repair may differ.

The role of parentinfant interactions

An important question is the extent to which the problems of infants with clefts may be explained by difculties in parentinfant interactions, since these predict poor child functioning (Eshel, Daelmans,

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health. Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA

116

Lynne Murray et al.

On the one hand, late repair may be advantageous because parents have time to adjust to the infants condition and plan for surgery (Munro, 1995), and parentinfant contacts immediately after the birth are not disrupted by the infants removal for surgery and post-operative recovery. On the other hand, in favour of early repair is the possible benet in terms of parentinfant interactions in the initial months (Bradbury & Hewison 1994; Munro, 1995), particularly since these involve face-to-face play, where parents of infants having early lip repair may nd it easier to respond to infant social cues. In turn, any enhanced responsiveness in parents of early repair infants may have benets for infant development. The only study examining this issue of which we are aware found no differences in the psychological impact of neonatal vs. late lip repair (Slade et al., 1999). However, the sample was small, and self-report, rather than observational measures, was used to assess maternal responses to the infant.

2. Cognitive decits in late repair group infants will be mediated by difculties in motherinfant interactions before surgical repair. 3. There will be no difference in cognitive or socioemotional functioning between infants having early cleft repair and control infants. In assessing the impact of cleft lip on mother infant interactions and infant development, a number of other factors require consideration. Thus, infant sex, family social class and maternal depression are all potentially important (Murray & Cooper, 2003). Furthermore, when considering results for the two index groups, a number of cleft-specic variables need to be taken into account: these include the severity of the cleft, i.e., whether cleft palate is also present, and the degree of visible deformity (Langlois, Ritter, Casey, & Swain, 1995). For later infant outcome, feeding method may be important, as well as hearing impairment and the experience of additional surgical procedures. In comparing the study groups, we therefore took these variables into account.

Aims of the current study

We aimed, rst, to assess the effects of cleft lip on infant cognitive and socio-emotional development at 18 months. Second, we aimed to assess the role of social interactions in the development of infants with clefts, and in particular, to determine whether interaction difculties might mediate any impact of the cleft on infant outcome at 18 months. We therefore assessed interactions at two, six and 12 months, and examined their relationship to infant outcome, as well as cleft status. Finally, we aimed to assess the effects of the timing of surgical lip repair; thus, two cleft groups were recruited: one where infants had neonatal lip surgery the early repair group; and a second where surgery was performed at 34 months postpartum the late repair group. We also recruited a normal control group. Given that previous work has been conducted with infants where the cleft has been unrepaired through the rst few postpartum months, we expected results for our late repair group to be in line with those in the literature, i.e., as showing poor cognitive, but not poor socio-emotional functioning. Moreover, and consistent with research in both normal and cleft populations, we expected the poor cognitive outcome for late repair infants to be accounted for by early parentinfant interaction difculties. For infants having early repair, by contrast, where there may be benets in terms of early motherinfant interactions, we did not expect such adverse outcome. Our specic hypotheses were as follows: 1. Infants having late repair will not differ from controls in their socio-emotional functioning, but will have poorer cognitive outcome than control infants.

Method
Sample
Index infants were recruited from four NHS regional centres for Cleft Lip and Palate Surgery. Two performed late lip repair (Radcliffe Inrmary, Oxford, and Alder Hey, Liverpool), and two, early repair (Stoke Mandeville, Aylesbury, and St Georges, London). Infants were admitted to centres by place of birth: those eligible for the study had isolated clefts of the lip (with or without cleft palate). Controls were recruited from maternity hospitals in the same regions (Royal Berkshire Hospital, Reading, Womens Hospital, Liverpool). Exclusion criteria for all infants concerned factors having independent effects on infant development, i.e., prematurity (<36 weeks gestation), low birthweight (<2.5 kg) or presence of other syndromes (e.g., oro-facial-digital syndromes, heart problems). In addition, if parents of index infants had elected to attend a centre other than their local one, they would not have been considered eligible for recruitment because of the self-selection involved: in fact, this did not arise. One hundred and forty-seven index infants were identied for recruitment by surgical staff when referred to the surgical team, prior to surgery. Parents were told that we were investigating the effects of cleft lip on parentinfant relationships and infant development: 110 (75%) consented, 28 (19%) refused, 4 (3%) could not be contacted and 5 (3%) were not approached. After each index infant was recruited, a potential control was selected by identifying an infant of the same sex whose time of birth was closest to that of the index infant, and, as far as possible, of the same birthweight and gestation. One hundred and twenty-ve potential controls were identied for recruitment on postnatal wards by research staff, who explained to parents that we wished to study motherinfant interactions and infant development in a normal population, for comparison with infants with

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health.

Effect of cleft lip on infant development

117

cleft lip: 100 (80%) consented, 22 (18%) refused, 2 (2%) could not be contacted, and 1 (1%) was ineligible. Subsequently, 5 further index infants were excluded as they were diagnosed with an additional disorder, and 2 dropped out, leaving 100 control and 103 index infants. Of the latter, 48 had early, neonatal, lip surgery, and 55 had late, 34 month, surgery. Ninety-six (96%) controls, 45 (93.8%) infants having early surgery and 49 (89.1%) having late surgery were assessed at 18 months. There were no differences between the study groups in rates of recruitment or follow-up. The study was approved by all local Ethics Committees. Parents gave written informed consent to participate.

Procedure
Research psychologists video-recorded motherinfant interactions in the home, in quiet conditions, at two, six and 12 months.1 For interactions at two and six months, the infant was placed in a supportive seat when alert and contented; the mother was asked to sit opposite her infant, and play with him/her for ve minutes. A mirror placed adjacent to the infant showed the mothers full-face reection, and the infants face and torso, in the same camera frame. At 12 months, mothers were videotaped when playing with their infants on the oor, using three age-appropriate toys, each presented for 2 1 minutes, following the proced2 ure described by Stein and colleagues (Stein, Woolley, Cooper, & Fairburn, 1994). At 18 months, trained psychologists, blind to the timing of lip repair, assessed infant cognitive development. Infant attachment to the mother was assessed, and mothers completed a questionnaire about infant behaviour problems. These assessments were conducted in University research rooms (Reading and Liverpool). At six months, index group mothers completed questionnaires evaluating the support received from the surgical teams to determine comparability between the centres. In addition, mothers were interviewed at two and six months, by psychologists trained to criterion, to establish whether they had been depressed since the birth. Medical details for index infants were obtained from the surgical teams and hospital records.

For the 12-month interactions, a further pair of researchers, who had no knowledge of the earlier interactions or maternal group, scored the videotapes, following the procedures of Stein and colleagues (1994) and Wolke and colleagues (1990). Measures were based on the literature on the effects of cleft lip/palate on motherinfant interactions, and that on predictors of infant cognitive and socio-emotional development. Three core measures (following Wolke, Skuse, & Mathison, 1990; and Murray, FioriCowley, Hooper, & Cooper, 1996), each rated on vepoint scales, were used at all ages, namely, maternal sensitivity (empathic, accepting, responsive and appropriate in relation to infant behaviour3, maternal positive involvement (emotionally expressive, positive, and actively engaged with the infant), and infant distress (fretful vocalisations, crying, frowning, grimacing, back-arching and squirming). Two further measures were used: for face-to-face play, each partners gaze to the others face was timed, and expressed as a percentage of the interaction time, using custom-made software to record gaze onsets and offsets in 20 millisecond intervals. For 12-month play, we rated maternal verbal elaboration on the toys (5-point scale). In line with our previous work (Murray et al., 1996), ratings were made for the whole ve minutes of face-to-face play; for 12-month interactions, ratings were made for play with each toy separately, and mean scores calculated. A random 10% sample of videotapes from each assessment was rated by two researchers to establish reliability. Intraclass correlations showed good agreement (range .71 to .94, mean .84.).

Rating procedures and measures Motherinfant interactions. Interactions were rated from videotapes by trained psychologists. For face-toface interactions, a masking disc was digitally superimposed on the mouth and nose of all infants, to ensure that ratings were done blind to infant group. 2 Different researchers scored maternal and infant behaviours.
1

While observations were generally conducted in the home, some were conducted in clinic/university rooms (20%). We examined the effects of location on motherinfant interactions: none was identied and this variable was not, therefore, included in subsequent analyses. 2 Inspection of the 12-months tapes, which comprised a wideangled camera view of the general play area, rather than focusing on participants facial communication, indicated that repaired clefts were not evident, and therefore masking was unnecessary.

Eighteen-month infant outcomes. Infant cognitive development was assessed using the Bayley Scales of Infant Development (1993) to obtain the Mental Development Index (MDI) score. This is a standardised, robust measure of infant cognitive functioning, widely used in research and clinical settings. It is reliable and, when used at this age, predicts later child intelligence quotient (Slater, 1995). Infant attachment was assessed using the Strange Situation Procedure (Ainsworth, Blehar, Waters, & Wall, 1978). Infants were lmed through a one-way mirror in an unfamiliar play-room for 21 minutes, both in the presence of their mother and of a female stranger. Infant responses to separation from, and reunion with, the mother were used to classify them as either secure (showing distress at the mothers departure, but seeking proximity and being comforted by her upon her return), or else insecurely attached (i.e., either showing avoidant, resistant or disorganised responses), following the criteria of Ainsworth and colleagues (1978) and Main and Solomon (1990). The videotapes were scored by two researchers, trained to reliability, who were blind to the nature of the study.
When rating maternal sensitivity, infant behaviour needs to be considered; e.g., maternal smiling or laughing may contribute to a rating of high sensitivity if occurring in response to excited, positive infant behaviour. However, if this maternal behaviour occurred in response to infant distress, it would be rated as low in sensitivity, since a sensitive response would be to show concern and empathy.
3

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health.

118

Lynne Murray et al.

Infant behaviour problems were assessed by maternal report. Mothers completed an age-modied version (Murray, 1992) of the Behaviour Screening Questionnaire (BSQ) (Richman & Graham, 1971). This is a widely used, reliable and valid measure that covers the frequency and severity of common childhood behaviour problems (e.g., sleeping and feeding difculties, separation problems). Total scores were used to characterise infant behavioural difculties.

Maternal interviews/questionnaires. Maternal depression was assessed by interview, using the Structured Clinical Interview for DSM diagnoses. This is a widely used, standardised instrument, with good reliability and validity (First, Spitzer, Williams, & Gibbon 1994). Professional support was assessed by index group mothers completing a questionnaire designed for the study, covering the personal (trust, conding fears), and practical (adequate information, and practical help) support provided by the surgical teams, using eight-point scales; Cronbachs alphas for the two measures were .89 and .66, respectively (see Hill, Murray, Woodall, Parmar, & Hentges, 2004 for further details). Cleft-specic factors. The following were recorded from medical records using a standard protocol, and were agreed upon by two of the study investigators: cleft palate, antenatal diagnosis, hearing problems (e.g., otitis media), surgical operations in addition to those for initial lip (and any palate) repair. In addition, whether or not the infants disgurement was severe was rated from standardised medical photographs taken before surgery (see appendix). Data analytic strategy
The three study groups (early repair, late repair and controls) were compared on infant 18-month outcomes, and on motherinfant interactions at two, six and 12 months. At two months, only the early index group had had lip repair, whereas by six months, both index groups had had surgery. One-way ANOVAs, ANCOVAs, KruskallWallis or Chi Square tests were used.4 Wherever these showed signicant effects, we conducted post-hoc tests of the pair-wise comparisons, making appropriate Bonferroni adjustments. Univariate associations were examined between the outcomes and the other relevant factors (SES, infant sex, etc.). If signicant, the relevant variable was included as a covariate in group comparisons. Where signicant differences emerged between the early and late surgery groups, we also examined the role of cleft-specic factors. In each case, we tested for main effects, and for whether cleft-specic variables moderated the effects of early and late lip repair. Furthermore, since the rate of antenatal diagnosis differed between the two index groups, we investigated its effects for all outcomes, regardless of whether or not early and late

repair groups differed. For normally distributed variables, two-way ANOVAs were conducted, with surgery group and cleft-specic variables as between-subject factors; for non-normally distributed variables, we conducted separate MannWhitney U-tests for early and late repair groups. Finally, where group differences in motherinfant interactions were found that paralleled differences in 18 month outcomes, and where these interaction variables were also associated with the infant outcome in question, we conducted mediational analyses (following Baron & Kenny, 1986), using hierarchical regressions to investigate whether the effects of group on infant outcome still obtained after taking account of mother infant interactions, or were reduced to non-signicance.

Results
The sample characteristics are shown in Table 1. There were no differences between the two index (early and later repair) and the control groups on any of the demographic variables, apart from maternal depression and feeding method at two months. Mothers in both index groups were more likely to be depressed than control mothers, and were more likely to bottle feed; differences between the early vs. late surgery group were not signicant. As to cleftspecic variables, only the rate of antenatal diagnosis distinguished the two surgery groups; this was higher in the early repair group.

Variable characteristics
Eighteen-month MDI scores were normally distributed, but BSQ scores were skewed. Motherinfant interaction variables were normally distributed, apart from maternal positive involvement at 12 months, and gaze; the former was normalised by square root transformation; maternal gaze was normalised by log transformations, but infant gaze could not be normalised.

Infant outcomes
Outcomes on the 18-month assessments are shown, by group, in Table 2. None of the potential covariates was associated with the socio-emotional outcomes, and therefore group comparisons were conducted unadjusted. There were no differences between the three groups in either behaviour problems (Kruskall Wallis, H (2) 2.072, ns), or quality of attachment, with all groups showing over two-thirds of infants to be securely attached (v2 (2) .302, ns). A contrasting picture emerged for infant cognitive functioning, where, after controlling for the effect of infant sex, a signicant difference between the groups was found on the Bayley MDI (F(2, 187) 5.573, p < .01); posthoc tests showed that infants having late lip repair performed signicantly more poorly than control

4 Developmental changes in social interactions across the rst year make longitudinal data difcult to interpret, and repeated measures analysis was not considered appropriate.

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health.

Effect of cleft lip on infant development Table 1 Sample characteristics Control (N 100) N Male Middle/upper SES 1st born Mothers depressed at 2 months at 6 months Breast feeding 2 months Mothers age (years) Birthweight (grams) Gestation (weeks) Cleft-specic factors Antenatal diagnosis Cleft lip only Severely disgured before lip surgery Hearing problems Extra operations Age at lip surgery (days) Personal support by surgical team 6 months Practical support by surgical team 6 months
a *

119

Early (N 48) N 32 24 18 11a*** 3 6a*** Mean 28.83 3406.24 39.49 N 23 21 24 12 4 Mean 4.83 Median 5 6 % 47.9 42.9 57.1 25.5 8.3 SD 3.62 IQR 4 4 % 65.3 49 36.7 22.9 6.3 12.5 SD 4.84 462.64 1.53 N 13b* 20 11 13 9 Mean 109.34b** Median 6 6 36 28 22 N

Late (N 55) % 64.3 50 39.3 14.5 5.7 17.3 SD 5.21 630.13 1.67 % 25 35.7 36.7 24.5 16.7 SD 36.06 IQR 5 4

% 65 55 44 2 10.4 49.4 SD 5.36 462.15 1.34

65 55 44 2 10 42 Mean 28.93 3528.97 39.88

8a* 3 9a*** Mean 29.22 3466.2 39.71

different from control group, bdifferent from early surgery group. p < .05; **p < .01; ***p < .001.

Table 2 Infant outcomes at 18 months Early (n 45) Mean Bayley MDI BSQ Insecure attachment
a b

Two-month interactions
Control (n 96) SD Mean SD

Late (n 49) Mean

a**b+

SD

93.04 11.97 87.37 3.89 3.37 3.24 N % N 13 31 10

15.86 94.88 11.5 3.67 3.23 2.97 % N % 27 25 32.1

different from control group. different from early surgery group. + p < .1; *p .05; **p .01.

group infants (difference 7.53, MDI points, around half a standard deviation, p < .01), whereas those having early repair were comparable to controls (difference 1.63 MDI points). Furthermore, when the two index groups were compared, there was a trend for the late repair group to have lower MDI scores (difference 5.9 MDI points, p .07). Antenatal diagnosis was not associated with infant outcome.

Motherinfant interactions
Means and standard deviations for each group on the different interaction variables at two, six and 12 months are shown in Table 3, together with specication of covariates included in the analyses.

Having adjusted for relevant covariates, the three groups were found to differ on all dimensions of motherinfant interactions. Post-hoc comparisons showed this was principally due to difculties in the late repair group: compared to controls, mothers of late repair infants were less positively involved and sensitive, and they looked less at their infants. In turn, infants in this group were more distressed, and they similarly looked less at their mothers. Early repair group motherinfant interactions were, by contrast, comparable to those of controls on all but the measure of maternal sensitivity, where they were rated more poorly. With regard to early vs. late group comparisons, mothers of infants having late repair were less positively involved with their infants, and their infants also spent less time looking at their mothers. Examination of cleft-specic factors showed infant disgurement to be important, with mothers of more disgured infants being less positively involved (F(1, 56) 5.58, p < .05); this effect was similarly pronounced in both surgery groups, the interaction between the two factors being non-signicant, (F(1, 56) .13, ns). In addition, although there was neither a main effect of severity of disgurement on infant gaze at two months (U 501, ns), nor an effect within the early surgery group (U 138, ns), for infants having late lip repair, the effect was signicant, with severely disgured infants looking at their mothers less than those who were not severely disgured (U 44, p < .05).

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health.

120

Lynne Murray et al.

Table 3 Motherinfant interactions at 2, 6 and 12 months Control Early Late Three group comparison Pair-wise post-hoc tests ControlEarly ControlLate EarlyLate

Mean 2 months Mother Pos. Involve Sensitivity Gaze Infant Distress Gaze 6 months Mother Pos. Involve Sensitivity Gaze Infant Distress Gaze 12 months Mother Pos. Involve Sensitivity Verbal elab. Infant Distress
a

SD

Mean N 42

SD

Mean N 45

SD

N 82 3.54 3.86 95.8 1.73 64.07 N 77 3.45 3.41 93.59 1.18 28.69 N 80 1.61 3.1 2.6 1.71 1.02 1.05 3.39 .92 31.89

3.24 3.13 95.28 2.05 57.94 N 40 3.40 3.05 94.7 1.73 35.61 N 41 2.16 2.97 2.56 1.80

1.04 .99 4.18 .95 32.81

2.68 2.81 91.74 2.34 36.51 N 40 3.03 3.43 94.24 1.11 23.39 N 40 1.80 2.9 2.64 1.71

1.06 .99 10.77 .91 29.99

F(2,159) 8.446a F(2, 159) 16.242a F(2,161) 3.723 F(2, 161) 6.463 K-W, H(2) 19.485

<.001 <.001 .026 .02 <.001

ns .001 ns ns ns

<.001 <.001 .027 .002 <.001

.023 ns ns ns .003

1.03 1.00 5.25 .87 20.59

1.24 1.20 4.51 .97 26.72

1.37 .81 5.7 .91 20.78

F(2,153) 1.876a F(2,151) 1.647a F(2, 139) .904a F(2, 140) .436 K-W, H(2) 5.476

ns ns ns ns ns

1.53 .74 .63 .51

2.17 .7 .71 .57

1.67 .74 .79 .61

F(2,156) .383 F(2,157) .9a F(2,156) .133 F(2,157) .438

ns ns ns ns

Positive involvement at 2 months controlled for depression; sensitivity at 2, 6 and 12 months, and positive involvement at 6 months, controlled for SES; maternal gaze at 6 months controlled for infant sex.

Infants with both cleft lip and palate looked less at their mothers in both index groups; this effect was signicant for the early repair group (U 132, p < .05), and was observable as a trend for the late repair group (U 156.5, p < .1). Antenatal diagnosis was not related to any twomonth interaction measure.

Six-month and 12-month interactions

There were no differences between the groups at either age on any of the interaction measures, and nor was there an effect of antenatal diagnosis.

Mediational analyses
Given that cognitive scores were signicantly lower in infants having late lip repair compared to control infants, and that all dimensions of two-month interactions were poorer in the late repair group, we next examined the relationship between these interaction variables and infant MDI scores, restricting analyses to the late repair and control groups. Each two-month motherinfant interaction measure was associated with 18-month infant cognitive functioning: mothers who were more sensitive, who looked at their infants more, and who showed high rates of positive involvement had infants with higher MDI scores (r .39, p < .001; r .20, p < .05; and r .26, p < .01, respectively); similarly, infants who were less distressed, and who looked more at their mother during the two-month interac-

tion, performed better (r . )18, p < .05, and r .25, p < .01, respectively). To determine whether the nature of two-month motherinfant interactions accounted for the difference in cognitive performance between late repair and control group infants, we conducted a series of multiple hierarchical regression analyses, in which each interaction variable was entered rst, before considering the effect of group. In each case, apart from that of maternal sensitivity, the effect of late repair group remained signicant when the mother infant interaction variable was taken into account. When surgery group was added to the model after having accounted for maternal sensitivity, its effect was no longer signicant (see Table 4). Thus, early maternal sensitivity mediated the effect of late lip repair on infant cognitive outcome. These results held when adjusting for the effect of infant sex. As a further check on the role of early interactions, we
Table 4 Hierarchical regression analysis: 2-month maternal sensitivity mediates effect of group (late repair vs. control) on infant MDI Predictor Step1 Constant Maternal sensitivity Step 2 Constant Maternal sensitivity Group (Late vs. Control) **p < .01; ***p < .001. B 76.26 4.83 80.79 3.97 )2.16 SE B 3.99 1.07 4.93 1.20 1.40 b D2 .15*** .387*** .017 .318** ).148

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health.

Effect of cleft lip on infant development

121

investigated whether two-month maternal sensitivity still predicted MDI scores, even when later sensitivity was taken into account. This proved to be the case: thus, although six and 12 month maternal sensitivity each predicted infant MDI (r .33, p < .001, and r .26, p < .01, respectively), the correlation between two-month sensitivity and MDI scores that controlled for both earlier measures was signicant (r .20, p < .05).

Discussion
In this study of infants with cleft lip (with and without cleft palate) and normal control infants, we found lowered cognitive scores in index infants having late lip repair. For infants having early, neonatal repair, by contrast, cognitive functioning was comparable to that of control infants. Index infants did not differ from controls in terms of attachment or behaviour problems, regardless of the timing of surgery. Motherinfant interactions at two months, but not six or 12 months, mirrored the pattern of group differences in infant cognitive outcome, with the late repair group showing more interaction difculties than controls on all dimensions. By contrast, the early repair group was broadly comparable to the controls. Further analyses showed that reduced sensitivity to the infant in late repair group mothers at two months mediated the adverse effect of late repair on infant cognitive outcome. Our study had a number of strengths: the sample size was relatively large, representative of the target population, and carefully recruited and monitored to ensure the inuence of additional syndromes or infant developmental problems did not confound the interpretation of the ndings. We also took account of other variables that might affect infant development. The fact that we found a pattern of development in those having late lip repair that was consistent with the results of previous studies of infants with unrepaired clefts, lends weight to previous ndings. Nevertheless, our nding that cognitive delay did not occur in infants having neonatal repair suggests that the overall conclusions regarding outcome for infants with clefts may require revision. Before drawing rm conclusions, however, two considerations are warranted. First, it is important to note that this is, to our knowledge, the only study to have investigated the effects of timing of lip repair on infant development, and our results require corroboration; this is particularly important in the context of multi-centre designs. Thus, while there was no evidence for centre-specic effects (rates of antenatal diagnosis did differ between the groups, but did not inuence any study outcome), and infants received either late or early repair solely on the basis of local hospital policy, rather than because parents elected for a particular surgical regime, it is possible that undetected differences

between centres may have inuenced the ndings. Second, although the Bayley MDI is quite robust, and when administered at this age predicts later child intelligence moderately well, it is important to establish the full clinical signicance of our ndings by determining whether group differences persist. Our nding that less sensitive maternal interactions accounted for the group differences in infant cognitive outcome is consistent with the general literature on the role of maternal behaviour in inuencing infant cognitive development, as well the ndings of Wasserman and Allen (1985) and Speltz and colleagues (2000) in relation to infants with clefts. Further, the fact that two-month interactions had an effect on infant MDI scores even when accounting for subsequent maternal behaviour is consistent with the view that there may be a sensitive period during which particular kinds of social experience may be important for normal cognitive development. Identifying such effects can be difcult: in normal samples, this is partly because the environment is often stable (Cicchetti, 2003; OConnor, 2003); and in high risk populations, although the early environmental disruption may be time-limited, it cannot always be assumed to occur independently of other factors (Rutter, 2000; OConnor, 2003). These difculties did not apply in the current study. Our ndings of interaction difculties in the late repair group raise the question of why these effects obtain. It could be argued that, prior to surgery, having a disgured infant may adversely affect maternal mental health, which in turn may affect interactions with the child. We assessed maternal depression in our study, but found no evidence for this explanation. Field and Vega-Lahr (1984) suggested that the disgurement to infants caused by unrepaired clefts not only makes these children less appealing to look at, but also makes it difcult for parents to interpret infant expressions. It was a limitation of our study that we could not examine interactions before two months, and so the way in which difculties developed, including the direction of effects, is unknown. Nevertheless, our ndings regarding the effects of initial infant disgurement within the late repair group do suggest that the appearance of the infant, which remained the same over the rst few weeks in this group, may have been important. Given that lip repair is generally conducted at 34 months, it is important that future research examine the natural history of mother infant relationships in the context of infant cleft lip, from birth through the early postpartum months, and explore the possible benets of interventions to facilitate social interactions.

Acknowledgements
This work was supported by the Medical Research Council, UK (G9629970). We thank Charlie Stanley,

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health.

122

Lynne Murray et al.

Polly Vittorini, Andreas Giannakakis, Konstantina Deligianni, Gillian Adams and Sarah Healy for assistance with coding videotapes, and conducting medical records searches.

Correspondence to
Lynne Murray, Winnicott Research Unit, School of Psychology, University of Reading, Reading RG6 6AL, UK; Email: lynne.murray@rdg.ac.uk

References
Ainsworth, M., Blehar, M., Waters, E., & Wall, S. (1978) Patterns of attachment: A psychological study of the Strange Situation. Mahwah, NJ: Erlbaum. Baron, R., & Kenny, D. (1986). The moderatormediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 11731182. Bayley, N. (1993) Bayley Scales of Infant Development (2nd edn, Manual). San Antonio, TX: The Psychological Corporation. Bradbury, E., & Hewison, J. (1994). Early parental adjustment to visible congenital disgurement. Child: Care, Health and Development, 20, 251266. Cicchetti, D. (2003). Editorial: Experiments of nature: Contributions to developmental theory. Development and Psychopathology, 15, 833835. Derijcke, A., Eerens, A., & Carels, E. (1996). The incidence of oral clefts: A review. British Journal of Oral and Maxillofacial Surgery, 34, 488494. Eshel, N., Daelmans, B., de Mello, M., & Martines, J. (2006). Responsive parenting: Interventions and outcomes. Bulletin of the World Health Organisation, 84, 991998. Field, T., & Vega-Lahr, N. (1984). Early interactions between infants with craniofacial anomalies and their mothers. Infant Behavior and Development, 7, 527530. First, M., Spitzer, R., Williams, J., & Gibbon, M. (1994) Structured Clinical Interview for DSM IV. New York: Biometrics Research. Goodacre, T., Hentges, F., Moss, T., Short, V., & Murray, L. (2004). Does repairing a cleft lip neonatally have any effect on the attractiveness of the subsequent repair? Cleft Palate-Craniofacial Journal, 41, 603608. Hill, J., Murray, L., Woodall, P., Parmar, B., & Hentges, F. (2004). Recalled relationships with parents and perceptions of professional support in mothers of infants treated for cleft lip. Attachment and Human Development, 6, 2130. Jocelyn, L., Penko, M., & Rode, H. (1996). Cognition, communication, and hearing in young children with cleft lip and palate and in control children: A longitudinal study. Pediatrics, 97, 529534. Kapp-Simon, K.A., & Krueckeberg, S. (2000). Mental development in infants with cleft lip and/or palate. Cleft Palate-Craniofacial Journal, 37, 6570. Koomen, H.M., & Hoeksma, J.B. (1993). Early hospitalization and disturbances of infant behaviour and

the motherinfant relationship. Journal of Child Psychology and Psychiatry, 34, 91734. Langlois, J., Ritter, J., Casey, L., & Swain, D. (1995). Infant attractiveness predicts maternal behaviors and attitudes. Developmental Psychology, 31, 464472. Main, M., & Solomon, J. (1990). Procedures for identifying infants as disorganised/disoriented during the Ainsworth Strange Situation. In M. Greenberg, D. Cicchetti, & E. Cummings (Eds.), Attachment in the preschool years (pp. 121160). Chicago, University of Chicago Press. Maris, C.L., Endriga, M.C., Speltz, M.L., Jones, K., & DeKlyen, M. (2000). Are infants with orofacial clefts at risk for insecure motherchild attachments? Cleft Palate-Craniofacial Journal, 37, 257265. Millard, T., & Richman, L.C. (2001). Different cleft conditions, facial appearance, and speech: Relationship to psychological variables. Cleft Palate-Craniofacial Journal, 38, 6875. Munro, I.R. (1995). A description of craniofacial anomalies: The mechanism & rationale of surgery. In R. Eder (Ed.), Craniofacial anomalies, psychological perspectives. New York: Springer-Verlag. Murray, L. (1992). The impact of postnatal depression on infant development. Journal of Child Psychology and Psychiatry, 33, 543561. Murray, L., & Cooper, P. (2003) Intergenerational transmission of affective and cognitive processes associated with depression: Infancy and the preschool years. In I.M. Goodyer (Ed.), Unipolar depression: A lifespan perspective (pp. 1746). Oxford: Oxford University Press. Murray, L., Fiori-Cowley, A., Hooper, R., & Cooper, P. (1996). The impact of postnatal depression and associated adversity on early motherinfant interactions and later infant outcome. Child Development, 67, 25122526. OConnor, T.G. (2003). Natural experiments to study the effects of early experience: Progress and limitations. Development and Psychopathology, 15, 837852. Richman, N., & Graham, P. (1971). A behavioural screening questionnaire for use with three-year-old children. Journal of Child Psychology and Psychiatry, 12, 533. Rutter, M. (2000). Psychological inuences: Critiques, ndings, and research needs. Development and Psychopathology, 12, 375405. Slade, P., Emerson, D.J.M.., & Freedlander, E. (1999). A longitudinal comparison of the psychological impact on mothers of neonatal and 3 month repair of cleft lip. British Journal of Plastic Surgery, 52, 15. Slater, A. (1995). Individual differences in infancy and later IQ. Journal of Child Psychology and Psychiatry, 36, 69112. Speltz, M., Endriga, M., Fisher, P., & Mason, C. (1997). Early predictors of attachment in infants with cleft lip and/or palate. Child Development, 68, 1225. Speltz, M., Endriga, M., Hill, S., Maris, C., Jones, K., & Omnell, M. (2000). Brief report: Cognitive and psychomotor development of infants with oral-facial clefts. Journal of Paediatric Psychology, 25, 185190. Speltz, M., Goodell, E., Endriga, M.., & Clarren, S.. (1994). Feeding interactions of infants with unrepaired cleft lip and/or palate. Infant Behavior and Development, 17, 131140.

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health.

Effect of cleft lip on infant development

123

Starr, P., Chinsky, R., Canter, H., & Meier, J. (1977). Mental, motor, and social behavior of infants with cleft lip and/or cleft palate. Cleft Palate Journal, 14, 140146. Stein, A., Woolley, H., Cooper, S., & Fairburn, C. (1994). An observational study of mothers with eating disorders and their infants. Journal of Child Psychology and Psychiatry, 35, 733748. Wasserman, G., & Allen, R. (1985). Maternal withdrawal from handicapped toddlers. Journal of Child Psychology and Psychiatry, 26, 381387. Wasserman, G., Lennon, M., Allen, R., & Shilansky, M. (1987). Contributors to attachment in normal physically handicapped infants. Journal of American Academy of Child and Adolescent Psychiatry, 26, 915. Wolke, D., Skuse, D., & Mathison, B. (1990). Behavioural style in failure to thrive infants: A preliminary communication. Journal of Pediatric Psychology, 15, 237253. Manuscript accepted 31 July 2007

Supplementary material
The following supplementary material is available for this article: Appendix. Ratings of severity of disgurement. This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/ j.1742-4658.2007.01833.x (This link will take you to the article abstract). Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

2007 The Authors Journal compilation 2007 Association for Child and Adolescent Mental Health.