4.1.1. Introduction

The aim of this review was to address the topic of identifying infants with weight loss that ought to raise concerns and at which time point in the first weeks it should be measured to best capture those infants that may need an intervention. To do this the report summarised the normal range of weight loss in the first weeks of life and when weight reaches its lowest point (nadir).

For full details see review protocol in Appendix D.

4.1.2. Description of clinical evidence

Nine studies (N=171,562) were included in the review.

Most of these studies focus on the first 3 to 4 days of life only (Bertini, 2015; Davanzo 2013; Flaherman 2010, ^{2013, 2015}; Martens 2007; Miller 2015). Two studies include weight loss measures beyond this follow-up, i.e. Macdonald 2003 up to 12 days, and Wright 2004 up to 6–8 weeks.

Due to differences in maximal weight loss, evidence was divided according to type of feeding:

• breastfed infants
• formula fed infants

• vaginal birth
• caesarean birth.

It was then determined what the time to maximal weight loss was and what percentage of weight loss was reported at this time point. We then reported the 50^th, 95^th and 97.5^th centile points at times reported in the studies.

Evidence from these studies is summarised in the clinical evidence profiles below (Table 11, Table 12, and Table 13). See also the study selection flow chart in Appendix F, summary charts in Appendix I, study evidence tables in Appendix G and exclusion list in Appendix H.

Table 11

Summary clinical evidence profile for weight loss in exclusively breastfed infants.

Table 12

Summary clinical evidence profile for weight loss in partially breastfed infants.

Table 13

Summary clinical evidence profile for weight loss in formula fed infants.

4.1.3. Summary of included studies

A summary of the included studies is presented in Table 10.

Table 10

Summary of included studies.

4.1.4. Clinical evidence profile

The clinical evidence profiles for this review question (normal weight loss in the first two weeks) are presented in Table 11, Table 12 and Table 13. These tables summarise results for each outcome across studies. When several studies reported findings for the same outcome, results were presented as ranges of percentages, hours or days. Quality was then rated for each study.

4.1.5. Economic evidence

As this question does not concern the competing uses of NHS resources it was not prioritised for health economic analysis. No health economic evidence was identified for this topic from the overall health economic search.

4.1.6. Clinical evidence statements

The timing of maximal weight loss (the weight nadir) was reported by 6 studies including 119,676 infants. Moderate quality evidence suggested this weight nadir is typically reached between 2 and 3 days after birth, regardless of method of delivery (vaginal birth versus Caesarean section) or feeding type (exclusively breast fed, partially breast fed or formula fed).

The maximal weight loss for exclusively breast fed infants was reported by 5 studies including 137,495 infants. Moderate quality evidence indicated a mean or median maximal weight loss ranging from 5.5% to 8.6%. The 95^th percentile for maximal weight loss ranged from 8.8% to 12.5%, and the 97.5^th percentile ranged from 9.4% to 13.8%.

The maximal weight loss for partially breast fed infants was reported by 5 studies including 49,747 infants. Moderate quality evidence indicated a mean or median maximal weight loss ranging from 5.5% to 6.3%. The 95^th percentile for maximal weight loss ranged from 9.5% to 12.0%, and the 97.5^th percentile ranged from 10.2% to 13.2%.

The maximal weight loss for formula fed infants was reported by 4 studies including 7915 infants. Moderate quality evidence indicated a mean or median maximal weight loss ranging from 2.4% to 7.5%. The 95^th percentile for maximal weight loss ranged from 6.3% to 11.6%, and the 97.5^th percentile ranged from 8.5% to 12.2%.

Low quality evidence from one study including 971 infants indicated that most had returned to their birthweight by 21 days. The 97.5th percentile for time taken to return to birthweight ranged from 16.7 days for formula fed infants to 21.0 days for exclusively breastfed infants.

4.1.7. Evidence to recommendation

The Committee agreed that the reviews for normal weight loss and adverse events related to weight loss thresholds are intrinsically the linked and one cannot be considered without the other in isolation. Evidence from both was discussed together to draft recommendations and therefore rationale for these is provided in section 4.3).

4.2.1. Introduction

This review aimed to determine the thresholds of weight loss in babies under 4 weeks that would lead to adverse events. For full details see review protocol in Appendix D.

4.2.2. Description of clinical evidence

Two studies were identified. One retrospective cohort study (N=874) from Taiwan was included in the review (Chang 2010). There were only available data for a threshold of 8% weight loss at 2 days after birth and a threshold of 11% at 3 days after birth. Evidence for these outcomes is summarised in Table 15.

Table 15

Summary clinical evidence profile for percent birth weight loss thresholds on 2^nd and 3^rd days of life to predict adverse outcomes in exclusively breastfed neonates.

One retrospective cohort study (N=1003) from Italy was also found (Davanzo 2013), although the design of the study did not directly satisfy the protocol criteria. However, it provided directly relevant data for the number of hypernatraemic infants above and below a threshold of 8% birth weight loss during the hospital stay immediately after birth. Evidence for this outcome is summarised in the clinical GRADE evidence profile below (Table 15).

See also the study selection flow chart in Appendix F, study evidence tables in Appendix G, full modified GRADE profile in Appendix J, forest plots in Appendix I and exclusion list in Appendix H.

4.2.3. Summary of included studies

A summary of the included studies is presented in Table 14.

Table 14

Summary of included studies.

4.2.4. Clinical evidence profile

The clinical evidence profile for this review question is presented is Table 15.

4.2.5. Economic evidence

As this question does not concern the competing uses of NHS resources it was not prioritised for health economic analysis. No health economic evidence was identified for this topic from the overall health economic search.

4.2.6. Clinical evidence statements

Low quality evidence from 1 retrospective cohort study (n=874) indicates that birth weight loss thresholds of 8% at two days after birth and 11% at three days after birth have low sensitivity for hyperbilirubineamia. Most infants with hyperbilirubineamia had lost less birth weight than these threshold values and would be missed if weight loss was used as the sole criterion for selecting those at risk of hyperbilirubinaemia.

Very low quality evidence from 1 retrospective cohort study (n=1003) suggests that a percentage birth weight loss below a threshold of 8% at any time during hospital stay immediately after birth has high sensitivity for hypernatraemia, although there is uncertainty around this estimate. According to this evidence a weight loss of less than 8% during hospital stay could be useful in ruling out hypernatraemia.

4.3.1. Relative value placed on the outcomes considered

The aim of this review was to address the topic of identifying infants with weight loss that ought to raise concerns and at which time point in the first weeks it should be measured to best capture those infants who may need intervention.

For normal limits of maximal weight loss, the main outcomes that the Committee considered for decision-making were the average weight loss observed in newborn babies in their first two weeks of life, the variation of this weight loss and the timing of maximal weight loss. Despite the outcomes were divided between breast fed, formula fed and breastfed partially infants, the Committee decided that the pattern of weight-loss seen in breastfed babies should be used as a guide for all babies, as it is explained in section 4.3.2. The Committee also discussed that it is important to know how long it would take infants to regain weight.

For percentage of weight loss associated with adverse outcomes, no evidence was retrieved for the outcomes listed as important and critical in the protocol. However, other adverse events often seen in neonates under 4 weeks old above and below a birth weight loss threshold of 10% such as hyperbilirubinemia and hypernatremia considered a proxy and taken into consideration for decision-making. Along with these, the Committee also took into account other adverse outcomes often seen in clinical practice.

4.3.2. Consideration of clinical benefits and harms

The Committee discussed the evidence that weight loss typically reaches its lowest point by 2 to 3 days after birth and 5% of breastfed babies lose 11% or more of their birth weight. The Committee agreed that if healthcare workers were aware of usual weight loss patterns after birth that they could reassure parents and avoid unnecessary interventions for the baby. Potential harms associated with unnecessary interventions include, parental anxiety, harms due to admission to hospital and harms due to supplementary feeding.

The Committee deliberated on the different thresholds of weight loss that were reported for breastfed and bottle fed babies. The evidence indicated that bottle fed babies initially lose less weight. This is perhaps to be expected given the normal volumes of colostrum in the early days compared to volumes of formula milk often offered. The pattern of weight-loss seen in breastfed babies should thus be used as a guide for all babies, as breastfeeding is the physiological norm. The Committee therefore agreed to set the same recommendations for all babies.

The Committee acknowledged that weight loss is usually due to body fluid shifts in the early days of life. If this was associated with clinical evidence of dehydration it would be pathological and a reason for intervention. The Committee also discussed that infants of mothers who received intravenous fluids before delivery or before caesarean section may show larger weight loss in the early days after birth. The large cohort studies included mothers who had received intravenous fluids but separate weight loss thresholds could not be extracted for the corresponding groups of babies. The Committee recognised that a larger fluid loss would be likely for these babies in the early days, but felt that the 10% threshold should still be an initial cause for concern. Intravenous fluid during labour may then be an issue that can be discussed as part of the clinical assessment. The Committee agreed that it would be important to evaluate an infant’s feeding as recommended if the weight loss was sufficient to raise some concern (more than 10%) and that the individual who observes the feeds has the relevant and appropriate expertise to do this (this could mean a health visitor or a trained person – usually with Baby Friendly Initiative accreditation). It was discussed that such expertise should not only be related to practical issues but should also include training in how to build positive relationships with mothers. Usually this assessment and observation would provide sufficient information to plan care for the infant but in some circumstances further investigations may be needed.

The infant who has lost more than 10% of their birth weight should be assessed for signs of effective feeding, milk transfer, urine and stool output. Healthcare professionals should also look for evidence of dehydration (because weight loss in the early days would usually be due to fluid loss). A related consideration is that there might be clinical evidence arising from the healthcare professional’s assessment that might point to an underlying cause such as an illness or disorder that might account for the weight loss. If such a disorder or condition is identified it would then lead to onward referral to an appropriate specialist who could then consider the relevant treatment options.

In relation to the time to regain weight the Committee agreed that the 3 weeks that were reported in the evidence were a good estimate for the time when weight should have returned to birthweight and the Committee agreed that both parents and healthcare professionals should be aware of this. If a baby has lost for example 10.1% and is otherwise doing fine parents and healthcare professionals may want to monitor whether the birthweight has returned by 3 weeks and if not consider further actions if necessary (such as relevant interventions and support).

The Committee recognised that weight loss is not the only indication of an unwell baby. They believed that healthcare professionals would understand what a clinical assessment should include, but wanted to particularly raise the issue of signs of dehydrations. These could include symptoms such as vomiting or diarrhoea. They thought it was important that healthcare professionals had a clear pathway to seek advice and medical or specialist feeding assessment if there were any concerns about weight loss or an apparently unwell infant. In this pathway an individualised approach to assessment was recognised by the Committee to be important. The Committee thought that such a pathway could prevent the need for admission while an assessment is carried out and interventions and support could be provided.

4.3.3. Consideration of economic benefits and harms

Any recommendation made in this area is likely to carry an indirect cost since identifying the thresholds of normal weight loss implies that some babies might have abnormal weight loss requiring treatment; this treatment is likely to carry an economic cost even if the baby is perfectly healthy. The alternative, failing to identify when babies do in fact need treatment, is likely to incur a significant financial and quality of life burden as these babies are unlikely to receive appropriate management and thus might present with more significant conditions later on in their life.

Committee opinion is that the indirect effects are likely to tend towards a saving to the NHS. In their opinion having a clearly identified threshold of usual weight loss will help caregivers identify when to suspect faltering growth, and health professionals will be able to reassure parents more effectively that lesser amount of weight loss are within the expected range. There is likely to be a quality of life improvement related to this cost saving, but there is no economic evidence comparing the magnitude of this quality of life improvement with the size of the cost saving, so it is not possible to determine which factor will ultimately be more important to the NHS.

In terms of highly indirect costs, there is evidence that continued breastfeeding can reduce overall healthcare spending by making certain illnesses less likely and promoting robust health generally. As this effect is ongoing over the lifetime of the child, it is likely that relatively small investments made in breastfeeding support early will be cost-effective given the accumulation of QALYs and costs offset over the lifetime of the child.

If there is any direct resource impact associated with the first of these recommendations, it is likely to be minimal; simply informing parents of the normal limits of weight loss and answering questions they might have. Consequently the resource impact of these recommendations will not be above the NICE ‘high impact’ threshold of £1 million per recommendation. Subsequent recommendations – such as that to carry out a clinical assessment - may carry a direct cost impact but it is thought unlikely that this would be ‘high impact’ as these assessments are already carried out in the NHS and the recommendations refine under what circumstances they should be offered.

4.3.4. Quality of evidence

The quality of the evidence about the normal limits of weight loss ranged from low to moderate, using the Munn 2014 quality checklist. The included studies typically used routinely collected measurements during hospital stay and detail about the method of weighing was lacking. After vaginal birth, mothers and babies were often discharged from hospital before maximum weight loss occurred. The largest studies were carried out in the USA, with potential demographic and maternity care differences to the UK population. Consequently, the Committee noted that there were some issues regarding indirectness of the setting.

The quality of the evidence about percentage of weight loss associated with adverse outcomes was very low to low as assessed by modified GRADE.

For the domain risk of bias, the studies were assigned ‘serious risk of bias’ since the outcome assessors were aware of group allocation and the design of the studies was retrospective. However, the Committee noted that the studies were well powered and that the time between delivery and evaluation of adverse outcomes was appropriate.

No serious issues were found regarding inconsistency (heterogeneity), only single studies were included.

Some issues regarding indirectness were also identified in this review question. The Committee discussed that the participants in one of the studies did not present with the 10% birth weight loss threshold and that the study was carried out in Japan, fact that raises concerns about applicability of the study to the UK setting. Along with this, the Committee noted that one of the studies did not control for confounding variables.

4.3.5. Other considerations

The Committee also discussed the need for consistent implementation of the recommendations into local pathways for assessment, support and referral when concerns about weight loss are raised.

It was agreed that early weight loss is not unusual and that therefore this is a distinct population in this guideline. Infants losing weight during these first few days (early days) would not be classified as having faltering growth. The Committee therefore agreed that these reviews and recommendations should be separated from the rest of the faltering growth guidance and that a preamble would be needed for this set of recommendations to highlight the difference between theses sections.

4.3.6. Key conclusions

Based on the available evidence the Committee extrapolated from indirect evidence of normal weight loss to make recommendations. Even though evidence for adverse events came from large data sets, it did not demonstrate an optimal weight loss threshold that identifies babies at risk of adverse outcomes. They therefore extrapolated from these reviews and their experience and expertise to provide a threshold (10% weight loss) that would not identify too many babies whilst capturing those where concerns would be justified. They also thought it important to highlight that commonly the birth weight is regained by 3 weeks of age.

Why this is important

Weight loss in breastfeeding infants in the first month of life can cause anxiety for parents and healthcare professionals. It can also incur costs to the NHS from admissions of the infant to hospital, with the potential for cessation of exclusive breastfeeding with its associated long-term health benefits.

• Practice varies across the UK. Robust evidence about which feeding interventions improve outcomes could inform practice, potentially reducing unnecessary and costly interventions and supporting parent-infant relationships and physical and emotional health.

Table 16

Research recommendation rationale.

Table 17

Research recommendation statements (PICO characteristics).