Dear Sir

Response to letter from James P Hoffman of Martek Biosciences Corp.

We completely disagree with Dr Hoffman's statement that "caregivers of preterm infants are not well served by our report of a 10 year follow-up of LCPUFA supplementation in preterm infants". Our study (1) presents the results of the longest follow-up of a randomised trial LCPUFA supplementation during infancy to date. We have explicitly acknowledged the shortcomings of the study (mainly cohort attrition) to a much greater extent than is typical in a study of this type, and we have discussed the limitations of our data and the need for further research into these outcomes. Our responses to Dr Hoffman's specific points are detailed in the Appendix.

It is worth noting here that, although the vast majority of infant formulas now contain LCPUFA, the scientific evidence base for their addition is recognised by most investigators and Key Opinion Leaders in the field to be weak; the most recent update of the Cochrane systematic reviews on LCPUFA supplementation of formulas for both preterm and term infants (encompassing 29 trials) concluded that there is no evidence for outcome benefits of the intervention, at least up to 18 months of age (2,3). We contend this field of research has been driven to an extent by enthusiasm and vested interest. As one of the major groups to do outcomes research in this area, we do not hold a fixed position but are open to the scientific evidence, and we have published on both positive and negative effects of supplementation in different trials. Our experience of publishing in this field has consistently been that publications supporting the addition of LCPUFA to infant formula are more readily accepted and less criticised than those which do not support the intervention, or which raise potential concerns. Thus studies such as that of Birch et al (4), on a small number of subjects, with significant attrition even in infancy but showing apparent large beneficial effects of LCPUFA supplemented formula on cognitive development have been widely cited as supporting the addition of LCPUFA. Indeed, Birch's study, which may have been one of the most influential trials driving the addition of LCPUFA to US formulas, was based on an incomplete follow up where only 19 subjects remained in the relevant intervention group, providing inadequate power to provide any realistic estimation of the treatment effect. It is odd then that our much larger study with more complete longer-term follow-up and a range of outcomes not previously examined should attract such critical comment, as that from Dr Hoffman. We have previously received criticism for other trials where we found potentially unfavourable effects of LCPUFA supplementation. For example, in one preterm trial we found preterm infants supplemented with LCPUFA had a long term reduction in linear growth (5) - yet another group that found the same thing but appeared nevertheless to favour single cell oil supplementation, received no such adverse comment (6). In contrast, whenever we have generated positive results, these have been accepted with enthusiasm. If our contention that there may be some underlying bias in this area, is true, this would not "well serve the caregivers of preterm infants" - or term infants - who are best served by objective reporting of scientific data in the interests of child health.

We note that we also measured cognitive outcomes during the follow-up of our current study and will be interested to see whether publication of these findings, which suggest some long term beneficial effects of LCPUFA supplementation, but in the same cohort with the same attrition rate and limitations, will attract the same level of scrutiny and criticism, which we doubt.

We hope that Readers will appraise our current manuscript in a critical, but importantly, open-minded manner, considering the limitations we have highlighted; and that eventually these data can be considered alongside those from other similar studies examine the long-term health effects of LCPUFA supplementation of infants formulas, in order to strengthen the evidence-base for future products.

Kathy Kennedy, Mary Fewtrell, Alan Lucas

Appendix. Response to Dr Hoffman's specific comments and questions:

1. We have not claimed in the paper that LCPUFA supplementation has programmed preterm girls to later obesity and hypertension! This was deliberate, so as not to cause alarm or extrapolate beyond our findings. As stated in our discussion, the girls in our study all had blood pressure and BMI currently within the normal range, and we merely speculated on the potential longer-term significance of our findings, for example, given data showing tracking of BP. 2. We have noted the greater height of the LCPUFA supplemented girls at follow-up in our paper, along with higher weight, skinfold thicknesses and fat mass etc. Fat mass and fat free mass adjusted for height (FMI and FFMI) were indeed not significantly different between groups, as reported in the manuscript. The difference in stature (and potentially more advanced pubertal development, as we have discussed) could explain the greater skinfold thickness and weight of the supplemented girls; however, it is equally possible that LCPUFA supplementation has resulted in greater fatness, which is recognised to be associated with increased stature and earlier pubertal development. We consider it inappropriate to adjust for height (an outcome measure potentially influenced by the intervention) in our main analyses. 3. Dr Hoffman asks for the baseline and follow-up characteristics of the girls who were followed up and those who were not seen to be presented; in the interests of brevity these results were not presented in the paper. However, we can confirm that girls who were seen were more preterm (30.4 wk vs 31.6 wks, p = 0.001), had lower birth weights (1379g vs 1500g, p = 0.05), spent longer in hospital (46 days vs 36 days, p = 0.013), were more likely to have required ventilation (52% vs 29%, p = 0.009), and their mothers were less likely to have been educated to degree level (0% vs 8%, p = 0.04) compared to those who did not take part in the follow-up. This will affect the generalisability of our findings to the original cohort. However, of greater relevance to the preservation of randomisation for those seen at follow-up, there were no significant differences between supplemented and control girls studied for baseline or follow-up characteristics, with the exception of the number of days of ventilation (4.4 days vs 1.1 days, p = 0.013). We did not record details of the timing of the introduction of solids nor of subsequent diet or physical activity in this study. The randomisation procedure should mean that these factors are equal in the two groups, although we accept that with attrition at follow-up, this may be questioned to some extent. 4. We do not 'ignore the importance of LCPUFA in preterm infant nutrition'; rather we question the evidence that adding LCPUFA to infant formulas in the manner used for formulas tested in clinical trials to date produces clinical benefit. The ESPGHAN Committee on Nutrition paper cited here itself acknowledges that 'the long-term effects on visual and neural development are not fully known', and the Expert group providing advice during the recast of the EU Directive on the composition of infant formulas in Europe (EFSA) concluded that there was insufficient evidence on which to make the addition of LCPUFA to infant formulas compulsory; the addition of LCPUFA to infant formulas is currently optional under EU regulations. A significant number of Key Opinion Leaders in this field have this view.

References 1. Kathy K, Ross S, Isaacs EB, Weaver LT, Singhal A, Lucas A, Fewtrell MS: Ten year follow-up of a randomised trial of long-chain polyunsaturated fatty acid supplementation in preterm infants: Effects on growth and blood pressure. Arch Dis Child 2010; 95: 855-595

2. Simmer K, Patole SK, Rao SC. Long-chain polyunsaturated fatty acid supplementation in infants born at term. Cochrane database of systematic reviews 2008; 23 (1) CD000376

3. Simmer K, Schulzke SM, Patole S. Long-chain polyunsaturated fatty acid supplementation in preterm infants. Cochrane database of systematic reviews 2008; Issue 1. Art No: CD000375. DOI: 10.1002/14651858.CD000375.pub3.

4. Birch E, Garfield S, Hoffman D, Uauy R, Birch D. A randomized controlled trial of early dietary supply of long-chain polyunsaturated fatty acids and mental development in term infants. Dev Med Child Neur 2000; 42: 174-181.

5. Fewtrell MS, Morley R, Abbott R, Singhal A, Isaacs E, Stephenson T, MacFadyen U, Lucas A. Double-blind, randomised trial of long-chain polyunsaturated fatty acid supplementation in formula fed to preterm infants. Pediatrics 2002; 110: 73-82.

6. Vanderhoof J, Gross S, Hegyi T. A Multicenter Long-Term Safety and Efficacy Trial of Preterm Formula Supplemented with Long-Chain Polyunsaturated Fatty Acids. JPGN 2000; 30: 121-127. Dear Sir

Reply to Dr Lapillone

We thank Dr Lapillone for his comments. We do not agree that our conclusions will confuse the readers of the Archives of Disease in Childhood, and this was certainly not our intention. The results are clearly presented (and summarised by Dr Lapillone). As he states, we assessed 'adiposity' in a number of ways - skinfold thicknesses, BMI, %fat and fat mass and fat free mass from deuterium dilution and normalised for height. We have clearly set out in the results and abstract which of these measures were significantly different between the supplemented and control groups; and in the abstract we do not claim that fat mass (or fat mass index) were significantly different.

Dr Lapillone is correct that our group has recommended using measures of fat and fat free mass normalised for height, rather than using % fat; however, this applies to whole body measurements. We have also advocated using skinfold thickness measurements as raw values to measure regional fat mass, and it is entirely plausible to have differences in regional adiposity not reflected in whole body measurements. Thus, we stand by our conclusion.

Kathy Kennedy, Mary Fewtrell, Alan Lucas

Conflict of Interest:

None declared

Editor:

Caregivers of preterm infants and children are not well served by the report of Kennedy et al.,[1] "Girls who were born preterm and received LCPUFA supplemented formula showed increased weight, adiposity and BP at 10 years, with potential consequences for later health." LCPUFA supplementation programming preterm girls to later obesity and hypertension is not supported by their data. The issue is height: supplemented girls averaged nearly 4 cm taller than controls with similar BMIs and fatness. "Differences in BP were not significant following adjustment for current weight." (All BPs were normal.) Failure to present analyses between girls of well-known confounders for weight such as time and type of introduction of solid foods, later childhood dietary intake and physical activity levels makes it difficult to comment on the effects of infant LCPUFA on weight at age 10. Baseline and follow-up characteristics are presented for all subjects, not specifically for girls, calling into question the validity of the girls' randomization at follow-up. Finally, the authors acknowledge[2] but seem to ignore their own admonition that a low follow-up rate (only 45%) is important when considering validity of conclusions and long-term consequences and obscure the critical role of LCPUFA in preterm infant nutrition.[3]

James P. Hoffman, MD Director, Medical Services Martek Biosciences Corp., Columbia, MD USA

[1]Kennedy K et al. The 10-year follow-up of a randomized trial of long-chain polyunsaturated fatty acid supplementation in preterm infants: effects on growth and blood pressure. Arch Dis Child 2010 95:588-595.

[2]Fewtrell MS et al. How much loss to follow-up is acceptable in long-term randomised trials and prospective studies? Arch Dis Child 2008;93:458-61.

[3]Agostoni C et al. ESPGHAN Committee on Nutrition. Enteral nutrient supply for preterm infants. Commentary from European Society of Paediatric Gastroenterology, Hepatology and Nutrition Committee on Nutrition. J Pediatr Gastroenterol Nutr 2010 Jan;50(1)85-91.

Conflict of Interest:

Employee of producer of LCPUFA supplements used in infant formula