Vitamin and trace element status and growth in children with short bowel syndrome being weaned off parenteral nutrition
Abstract
Objectives
Children with short bowel syndrome (SBS) are at increased risk of micronutrient deficiencies. We aimed to study weaning off parenteral nutrition (PN) as a risk factor for vitamin and trace element deficiencies and growth deficit.
Methods
Children born between 2010 and 2019, treated for SBS in Helsinki University Hospital, were followed for their nutrition, growth, and vitamin and trace element biochemical status. Information on these as well as patient characteristics and intestinal anatomy were collected from electronic patient records.
Results
A total of 59 patients with SBS were identified. The weaning period was found to be a risk factor for micronutrient deficiencies and growth. Most frequent deficiencies were detected for vitamin K as suggested by prothrombin time (PT) (91% of patients), vitamin D (70%), and vitamin A (67%). Other deficiencies were less common with 44% for vitamin B12, 25% for selenium, 19% for copper, 24% for zinc, and 11% for vitamin E. For all the studied micronutrients, 3 months before to 1 month postweaning had the highest rate of micronutrient deficiencies, except for vitamin B12, for which the highest proportion of deficiencies presented at 1–6 months postweaning.
Conclusions
The weaning period from PN is a risk for micronutrient deficiencies. Regular monitoring and sufficient supplementation are essential for preventing malnutrition.
Highlights
What is Known
-
Children with short bowel syndrome (SBS) are at increased risk for micronutrient deficiencies.
-
Practices on micronutrient supplementation vary.
What is New
-
Time of weaning off parenteral nutrition (PN), when only supplemental PN is provided and absorptive capacity is just developing, is a particularly vulnerable period for micronutrient deficiencies.
-
Vitamin B12 deficiency often develops later while on full enteral feeding, necessitating long-term surveillance in SBS patients.
1 INTRODUCTION
Intestinal failure (IF) is defined as the inability to meet the enteral nutrition (EN) needs for adequate growth and energy expenditure. Short bowel syndrome (SBS) causes the majority of IF seen in pediatric patients.1 Outcomes of SBS are related to remaining bowel anatomy which influences duration and amount of parenteral nutrition (PN), although enteral adaptation is a multifaceted process.2 Impaired absorption of vitamins and trace elements is prevalent in SBS in addition to that of macronutrients. Before routine micronutrient supplementation in PN, micronutrient deficiencies were common in children with SBS.3 Nowadays, vitamins and trace elements are added to parenteral mixtures to ensure adequate micronutrient balance.4, 5 Once weaning off PN to EN, the risk of vitamin and trace element deficiencies emerges due to decreasing parenteral supplementation and limited absorptive capacity. Indeed, previous studies have shown deficiencies in children with IF weaned or in the process of weaning off PN for trace elements such as zinc, copper, and selenium,6-12 and vitamins A, D, E, and B12.7, 11, 13-15 Micronutrient deficiencies may affect growth, immune responses, and intestinal mucosal functions, and cause metabolic disturbances, neurocognitive challenges, and other organ dysfunctions.16
Evidence on factors associated with postweaning nutritional deficiencies is scarce. We aimed to study the micronutrient status in children with SBS to explore potential risk periods or disease-specific details for each micronutrient.
2 METHODS
2.1 Study design
We conducted a retrospective chart review of patients with SBS, born between January 1, 2010, and December 31, 2019, treated at Helsinki University Children's Hospital. We defined SBS as 50% or less of the age-adjusted small bowel remaining or PN > 3 months.17 We excluded other etiologies of IF.
We collected information on patient characteristics, intestinal anatomy, growth, nutrition, and laboratory measurements from electronic patient records. Nutrition included PN and EN, oral nutritional supplements, and vitamin and trace element preparations administered orally, enterally, intramuscularly, or intravenously. We calculated energy intake from PN and the proportion of PN of resting energy expenditure (REE), as estimated by Schofield equations based on height and weight. We categorized patients as being on none, supplemental (<75% of REE), or predominant (>75% of REE) PN.18 Information was gathered 3 months before weaning off PN, at stopping PN, 6 months postweaning, and at the last available visit, and cross-sectionally at the age of 1 year. All available laboratory measurements of vitamins and trace elements were collected.
Trained nurses measured the weight and height of the children at each visit. Growth was analyzed according to Finnish National Growth Charts, using z-scores for height and percentiles (near equivalent to z-score) for weight-for-height/supine length.19
From 2011 onwards, the PN provision changed from individual compounded bags to factory-made two- or three-chamber bags and tailored lipid emulsions. Before 2011, lipid emulsions containing olive and soy oil were used, later a multisource lipid emulsion (SMOFLipid) has mainly been used in long-term home PN. Water- and fat-soluble vitamins and trace elements are routinely added according to the product stability information and patient needs: Soluvit® and Vitalipid® (Fresenius Kabi) are added to lipid infusion and Peditrace® (Fresenius Kabi) for younger, and Addaven® (Fresenius Kabi) or Cernevit® (Baxter Inc.) for older children to a multi-chamber bag. Details on our SBS care protocol have been published previously.20
During home PN and in the first year after weaning, the patients attend an outpatient clinic every 1–3 months. Dietetic follow-up is provided based on individual needs, annually at a minimum. Protocol-based laboratory tests are taken at every visit. Post PN, we use routine enteral fat-soluble vitamin supplementation (vitamins D, A, E, and K), but trace elements or other vitamins are only given if a deficiency develops or if dietetic assessment indicates suboptimal intakes. For children lacking ileum, vitamin B12 is administered orally or, if blood concentrations are low despite adequate oral supplementation, intramuscularly. To estimate vitamin K availability, we use prothrombin time (PT) in the absence of significant liver disease.
2.2 Biochemical assessment
Blood samples were collected routinely every 6–12 months from all children with SBS as part of their follow-up regime. Vitamin and trace element analyses were carried out at the Helsinki University Hospital laboratory with standard protocols.
2.3 Statistical methods
Data are described as medians and interquartile range (IQR) or mean (standard deviation [SD]) as appropriate. Comparisons between groups at specific time points were made using one-way analysis of variance for normally distributed data and Kruskal–Wallis for nonparametric data. Comparisons between groups were tested with an unpaired t-test for normally distributed and by the Mann–Whitney U test for nonparametric data. We measured correlations using the Pearson correlation coefficient for normal and the Spearman coefficient for nonnormally distributed data. Normality was tested by Pearson test and visual inspection. Time from weaning was categorized as more than 3 months before, 3 months before to 1 month post, 1 month post to 6 months post, and more than 6 months postweaning, or more than 3 months before, 3 months before to 3 months post, and more than 3 months postweaning, depending on the statistical model. Generalized estimating equations were used to find associations between micronutrient blood concentrations and background factors. IBM SPSS 27 (IBM Corp.) was used for statistical analysis. p-Value <0.05 was considered statistically significant.
2.4 Ethical board approval
An institutional research permit was granted. Ethical committee approval was not required as the study was based on chart review.
3 RESULTS
3.1 Patients
We identified 59 patients with SBS. The underlying cause for SBS and other patient characteristics are described in Table 1. At the age of 1 year, 31 (59.6%) thrived on full EN, 6 (11.5%) received supplemental PN, and 15 (28.8%) received predominantly PN. At the last follow-up, with a median age of 4.5 years (range 3 months to 10 years), 12 (20.3%) children received predominantly PN, 5 (9.1%) received supplemental PN, and 38 (69.1%) children thrived on full EN; 3 (5.4%) were lost to follow-up. Twelve (20.3%) children received tube feeds.
| Sex, males n (%) | 31 (52.5%) |
| Gestational age, weeks, median (IQR) | 31 (25–35) |
| Underlying cause for SBS, n (%) | |
| NEC | 31 (52.5%) |
| SBA and multiple atresia | 12 (20.3%) |
| Volvulus and midgut volvulus | 7 (11.9%) |
| Gastroschisis and combined SBA and gastroschisis | 4 (6.8%) |
| Hirschsprung's diseasea | 4 (6.8%) |
| PIPOa | 1 (1.7%) |
| Small bowel at shortest, % median (IQR) | 42 (27–67) |
| Ileocecal valve, yes, n (%) | 33 (55.9%) |
| Functional circuit at last follow-up, n (%) | |
| Jejunum | 7 (11.9) |
| Jejunum, colon | 18 (30.5%) |
| Jejunum, ileum, colon | 34 (57.6%) |
| PN duration in those who were weaned, months, median (IQR) | 5.1 (2.6–10.9) |
| Weaned off PN, yes, n (%) | 39 (66.1%) |
- Abbreviations: IQR, interquartile range; NEC, necrotizing enterocolitis; PIPO, pediatric intestinal pseudo-obstruction; PN, parenteral nutrition; SBA, small bowel atresia; SBS, short bowel disease.
- a Children also had short bowels.
3.2 Growth
Growth was affected by bowel anatomy and the mode of feeding. In univariate analysis considering the child's overall growth, children with small bowel length ≤26 cm, and children with no ileum or ileocecal valve had lower height SD than the children with small bowel length >26 cm or those with ileum in continuity and ileocecal valve present (p < 0.05 for all). In addition, children receiving supplemental PN had shorter height SD than those on full enteral or predominantly parenteral feeding (−3.0 SD, −1.3 SD, and −1.4 SD, respectively; p = 0.04). However, in a multivariate model including functional bowel circuit, the use of hydrolyzed formula, and feeding modality, only bowel anatomy was associated with height: mean adjusted height z-scores were −3.2 for children with jejunum only, −1.5 for children with jejunum and colon in continuity, and −0.9 for children with jejunum, ileum, and colon in continuity (p = 0.010); PN was no longer a statistically significant determinant of height. For weight-for-height, no explanatory factors were identified.
3.3 Enteral nutrition
At the age of 1 year, 52 children were evaluated. Of the children, 40 (67.8%) ate solid foods and 34 (65.4%) received a hydrolyzed formula (Table S1). Type of enteral feed or diet and provision of vitamin or trace element supplementation were similar between groups receiving full enteral or supplemental or predominantly PN. Exceptions included breastfeeding, which was most common in those on predominantly PN (p = 0.045), the use of hydrolyzed feed, which was most common in children receiving supplemental or predominantly PN (p = 0.027) and the use of intramuscular (i.m.) vitamin D preparation, which was highest in those receiving supplemental PN (p = 0.007). Similarly, hydrolyzed infant formula and enteral tube feeds were most common in those receiving supplemental or predominant PN at the last follow-up (Table 2). The use of vitamin and trace element supplementation is described in Table S2.
| All patients, n = 55 | Full EN, n = 38 (69.1%) | Partial PN, n = 5 (9.1%) | Predominantly PN, n = 12 (21.8%) | p for difference | |
|---|---|---|---|---|---|
| Age, years, mean (SD) | 4.5 (3.0) | 4.8 (2.9) | 6.2 (4.5) | 3.0 (1.9) | 0.120 |
| Height SD, mean (SD) | −1.4 (1.6) | −1.3 (1.5) | −3.0 (2.8) | −1.4 (1.1) | 0.040 |
| Weight-for-height SD, mean (SD) | −0.4 (0.6) | −0.4 (0.7) | 0.07 (1.3) | −0.7 (0.4) | 0.224 |
| EN, n (%) | |||||
| Eating solids | 51 (92.7) | 37 (97.4) | 4 (80.0) | 10 (83.3) | 0.103 |
| Breast milk | 0 | 0 | 0 | 0 | |
| Standard infant formula | 3 (5.5) | 1 (2.6) | 0 | 2 (16.7) | 0.223 |
| Hydrolyzed formula with MCT | 11 (20.0) | 4 (10.5) | 2 (40.0) | 5 (33.3) | 0.025 |
| Enteral tube feeds | 12 (21.8) | 5 (13.2) | 2 (40.0) | 5 (41.7) | 0.047 |
| Sip feeds | 5 (9.1) | 4 (10.2) | 0 | 1 (8.3) | 1.000 |
| MCT oil | 3 (5.5) | 3 (7.9) | 0 | 0 | 1.000 |
| Baby rice mixed with feed | 3 (5.5) | 1 (2.6) | 0 | 2 (16.7) | 0.223 |
| Protein supplement | 1 (1.8) | 0 | 0 | 1 (8.3) | 0.309 |
- Abbreviations: EN, enteral nutrition; MCT, medium chain triglycerides; PN, parenteral nutrition; SD, standard deviation.
3.4 Biochemical assessment
Deficiencies of micronutrients were common (Table 3). In biochemical assessment, low concentrations of vitamins were more common than those of trace elements with most patients having at least one suboptimal measurement of vitamin A (n = 39, 67.2%), vitamin D (n = 39, 70.0%) or PT (n = 53, 91.4%). Deficiency of vitamins E and B12 were less common, in 6 (10.5%) and 20 (44.4%) children, respectively. Low levels of zinc were seen in 11 (24.4%), of selenium in 13 (25.0%), and of copper in 6 (18.8%) children at least once during the study period. Most children had suboptimal vitamin levels for multiple vitamins at least once during the study; 6 (10.2%) children had low status in one, 19 (33.3%) in two, 18 (31.6%) in three, 11 (19.3%) in four and 3 (5.3%) in all five vitamins assessed. Any of the three trace elements was measured at least once in 55 children. Of them, 20 (36.3%) had suboptimal levels in one and 5 (9.1%) in two of the trace elements; none had low levels in all three.
| Measured in how many children | Deficiency at least once | Estimated proportion of children with deficiency | |||||
|---|---|---|---|---|---|---|---|
| <3 months before weaning | −3 to +0.99 months from weaning | +1 to 5.99 months from weaning | ≥6 months from weaning | p | |||
| A | 57 | 39 (67.2%) | 33% | 37% | 25% | 7% | <0.001 |
| E | 57 | 6 (10.5%) | 1% | 8% | 5% | 1% | <0.001 |
| D | 55 | 39 (70.0%) | 30% | 58% | 25% | 9% | <0.001 |
| PTb | 58 | 53 (91.4%) | 66% | 66% | 25% | 21% | <0.001 |
| B12 | 45 | 20 (44.4%) | 6% | 13% | 44% | 19% | <0.001 |
| Zn | 45 | 11 (24.4%) | 5% | 12%c | 9%d | 0.554 | |
| Se | 52 | 13 (25.0%) | 10% | 20% | 18% | 5% | <0.001 |
- Abbreviations: PT, prothrombin time; Se, selenium; Zn, zinc.
- a Vitamin and trace element deficiency at least once during the study and the estimated proportions of low blood concentrations at different stages of weaning, taking into account repeated measures. Copper measurements were too few to perform the analysis. Reference values defined by Helsinki University Hospital laboratory are: vitamin A, 0.7–1.5 (1–6 years), 0.9–1.7 (7–12 years), 0.9–2.5 (13–19 years) µmol/L; vitamin E 7–21 (1–6 years), 10–21 (7–12 years), 13–24 (13–19 years) µmol/L; PT 70%–130%; vitamin D > 50 nmol/L; vitamin B12 > 35 pmol/L; zinc 7.7–15 µmol/L; selenium 0.45–0.9 (<1 year), 0.41–1.07 (1–4 years), 0.52–0.9 (5–9 years), 0.51–1.04 (1–16 years); copper 7.9–20.5 (7–12 months), 12.6–23.6 (1–6 years), 13.2–21.4 (6–10 years), 12.6–19.0 (10–14 years), and 10.1–18.4 (15–19 years).
- b PT (in the absence of liver disease) was used as a marker of vitamin K status.
- c −3 to +2.99 months from weaning.
- d ≥3 months from weaning; due to small number of samples, division into four categories was not possible.
At the age of 1 year, biochemical assessment of vitamins and trace elements did not reveal differences between feeding modalities (data not shown). The number of each vitamin or trace element in blood samples varied (Table 3).
In a pairwise analysis with pre- and postweaning blood concentrations ±24 weeks from weaning, serum vitamin A concentrations (median 0.87 vs. 1.20 µmol/L, p < 0.001, n = 24), PT (median 51.3% vs. 66.3%, p < 0.001, n = 26), and selenium concentrations (0.55 vs. 0.83 µmol/L, p = 0.005, n = 12) inclined from before to after weaning off PN. No statistically significant changes were seen in biochemical assessments of vitamin E (n = 25), vitamin D (n = 17), or zinc (n = 5) in the pairwise comparison.
In a multivariate analysis adjusted for bowel length and anatomy, ileocecal valve, child's age and gender, vitamin A levels were higher in those on full EN compared to those receiving PN with adjusted mean (confidence interval, CI) levels of 1.17 (1.07–1.27) versus 0.83 (0.76–0.89), respectively (p < 0.001). On the contrary, vitamin E levels were lower in those on full EN compared to those receiving PN with adjusted mean (CI) levels of 22.8 (21.0–24.7) versus 25.0 (23.4–26.8), respectively (p = 0.001). Univariate analysis showed some associations between bowel anatomy and vitamin and trace element blood concentrations (Table S3); however, these associations did not remain in multivariate analysis after adjusting for nutrition and age of the child (Tables S4 and S5).
We also analyzed whether the period of weaning off PN (from 3 months before to 1–3 months post weaning) is a risk for suboptimal vitamin or trace element status. Blood concentrations were lower during the weaning period for vitamin E (p = 0.001) and vitamin D (p < 0.001), as well as for zinc (p < 0.001) and copper (p < 0.001) (Tables S4 and S5). Micronutrient deficiencies within the weaning period were frequent for vitamin D (58%), vitamin A (37%), and vitamin E (8%), and for selenium (20%) and zinc (12%), whereas low vitamin B12 concentrations were seen most often (in 44% of the patients) from 1 to 6 months postweaning (Table 3). Copper deficiency was more common before weaning, in 23% of the patients.
4 DISCUSSION
We found that most children with SBS had suboptimal vitamin and trace element status at some point during their treatment. The highest proportion of children with micronutrient deficiencies occurred during weaning off PN.
Vitamin A levels were lower in children on PN than in children weaned off PN. Other nutrients with a risk of suboptimal status before weaning off PN were PT as a marker of vitamin K status and selenium. In contrast, vitamin E concentrations were higher in children while on PN than postweaning. Blood concentrations of vitamins E and D, and zinc and selenium tend to be the lowest during the weaning period. Vitamin B12 deficiency tended to become apparent only after weaning off PN.
Although two other studies have demonstrated micronutrient deficiencies during the weaning period, the risk of specific micronutrient deficiencies varies,12, 21 possibly reflecting differences in micronutrient supplementation practices. Similarly to Namjoshi et al.,10 we found that having any micronutrient deficiency was common: in our study, approximately 70%–90% of children had had low blood concentrations of vitamin A, vitamin D, or vitamin K as indicated by PT, as opposed to 90% of children in Namjoshi's study in children on chronic PN. Similarly to their study, we could not find associations between age, time on PN, or sex of the child with trace elements. Namjoshi et al. found an association between high body mass index and multiple micronutrient deficiencies; in our study, weight-for-height SD was not associated with micronutrient status.
When adjusted for confounders, we found no meaningful associations between bowel anatomy and blood micronutrient concentrations, likely due to our rigorous follow-up protocol and prompt supplementation practice.
Opposed to the 23% seen in our study, Adler et al. reported a significant proportion of infants with an ostomy and PN with copper deficiency22; however, their study was conducted in a neonatal intensive care unit and therefore serum levels may have been falsely low due to systemic inflammatory response23 and thus the populations may not be comparable.
Feng et al. described 97% of children having at least one micronutrient deficiency during PN, and 83.9% after weaning into EN.7 Similarly, Ubesie et al. reported 56% of children with at least one micronutrient deficiency during the transition from PN to EN11 and Young et al. found 33% with at least one vitamin and 77% with at least one trace element deficiency during PN; deficiencies of vitamins were increased to 70% while deficiencies of trace elements stayed at 77%.12 In our population, the highest proportions of low blood concentrations for vitamins and trace elements were from 3 months before to 1 month postweaning off PN. Differences in defining weaning from PN or differing populations and supplementation and eating habits may explain the differences.
Like Ubesie et al.,21 we found better vitamin A status after weaning off PN than before, while other studies have noted high risk of vitamin A deficiency after weaning.12, 24, 25 Parenteral vitamin supplementation is decreased in proportion with the proportion of energy derived from PN, so vitamin A supplementation altogether has been lower before than after weaning, when oral supplementation is more common (Table S2). Unfortunately, we do not have detailed time-stamped records of PN vitamin supplementation, so we cannot calculate the exact supplemental intakes at each time point. Vitamin K deficiency, as indicated by PT, was more common in our study compared to a previous report with normal coagulation times in children with SBS.26
Active absorption of vitamin B12 occurs in the terminal ileum. After ileal resection, intestinal adaptation sufficient enough for vitamin B12 absorption may take years.14 Accordingly, we found the highest proportion of children with low vitamin B12 concentrations +1 to 5.99 months from weaning. In our population, vitamin B12 deficiency was not associated with bowel anatomy. This is likely due to our rigorous follow-up and supplementation practices. Vitamin B12 supplementation was rare (Table S2), however due to the retrospective nature of the study, it is possible that patient records are missing some vitamin injections, as these may have been given in local hospitals.
Some centers withdraw copper from PN solutions in cholestasis since copper excretion has been suggested to be impaired, but withdrawal may lead to clinical copper deficiency.9, 27, 28 In our center, withdrawal is not customary and copper deficiency was less common than in other studies.22, 27
Zinc absorption in infants with ostomies is approximately 10 percent of intake both prior and post reconnective surgery,6 underlining the limited absorptive capacity. Another study in children with IF found plasma zinc levels correlating negatively with stoma output.29 We did not find associations between zinc concentrations and bowel anatomy in our multivariate models, which could be due to sufficient supplementation, or a small sample size.
Since increasing vitamin D fortification of milk products in 2003, vitamin D levels in the general Finnish population have increased from 48 to 65 nmol/L.30 In a publication from 2013 by our group, we demonstrated that 41% of children had suboptimal vitamin D serum levels during or after weaning off PN13 and similarly vitamin D deficiency has been common in other studies.10, 31, 32 In the present study, the proportion of children with suboptimal concentrations was 58% during the weaning period. This is reflected in the high use of i.m. vitamin D injections, in 67% of the children at the age of 1 year receiving supplemental PN.
We demonstrate that the weaning period is not only reflected in suboptimal micronutrient biochemical status but also temporarily in growth: children receiving supplemental PN, indicating weaning off PN, also have shorter height SDs than those on full enteral or PN. We cannot rule out that this finding is not due to overfeeding during PN; however, weight-for-height does not support overfeeding. A previous study from our center also showed that low levels of essential fatty acids are common during the weaning period,18 highlighting the vulnerability of the rehabilitation process. Tube feeding is also associated with poorer growth, but this is not considered causal; rather, children with poor growth are more likely to be tube-fed in an attempt to enhance their nutritional status.
Mode and type of feeding changes over time. The use of tube feeding and hydrolyzed formula is more common at the age of 1 year than at the last follow-up (Tables 2 and S1). Whether a hydrolyzed formula is necessary for the growth and well-being of children with SBS can be questioned,33 and this practice needs to be evaluated. The low rate of children eating solid foods at the age of 1 year, 77% of the children, also prompts action in preventing feeding difficulties in infants on PN.34
A strength of our study is the number of children with follow-up data, and centralizing the patients on PN into one center, so the treatment protocol is identical. A limitation of this study is the retrospective nature and the erratic biochemical assessment. In addition, vitamin intakes from PN, EN, and food were not quantified. In addition, C-reactive protein was not measured together with micronutrients. Thus, the effects of the systemic inflammatory response cannot be ruled out as a cause of low micronutrient blood concentrations. Furthermore, the last follow-up was at different ages due to the cut-off in the study period; children were followed for varying lengths of time. A follow-up study with a longer following period in the future will help understand the caveats of the weaning period with more children being weaned off PN with time.
In conclusion, micronutrient deficiencies are common in children on PN and in the process of weaning off PN and shortly after weaning, which stresses the importance of regular monitoring and adequate supplementation.
ACKNOWLEDGMENTS
The authors are grateful to the families that participated in the study. The authors also thank biostatistician Tuija Poussa for statistical analyses using Generalized Estimating Equations (Tables S4 and S5). The funding was provided by the Helsinki and Uusimaa Hospital District Research Fund.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
