Materials and methods

We conducted a retrospective chart review of subjects < 18 years of age who underwent CHD surgery at Children’s Hospital, New Orleans between 2018 and 2022. The range of cardiac procedures patients underwent included Tetralogy of Fallot, Bidirectional Glenn, Fontan, and Norwood procedures, as well as aortic arch, right ventricular outflow tract, left to right shunt, and anomalous pulmonary venous return repairs. Each case was identified as a subject with post-operative diagnosis of chylothorax. Each case was matched 1:1, by procedure and age, to a control who received CHD surgery at Children’s Hospital, New Orleans but did not develop chylothorax. A system for classifying the risk of death associated with CHD surgery developed by the Society of Thoracic Surgeons – European Association for Cardio-Thoracic Surgery (STAT Mortality Category score) was calculated for each case based on procedure types to determine whether there were any differences in mortality risk between the two groups.

After obtaining study approval from our internal institutional review board, we utilised our electronic medical record system, EPIC, to identify cases using relevant International Classification of Diseases codes (ICD-10 codes: J94.0 (chylous effusion), I89.8 (chylothorax), and J90.0 (pleural effusion)). Each chylothorax diagnosis was confirmed with a pleural fluid triglyceride level ≥ 110 mg/dL. A participant flow chart outlines the screening and inclusion criteria in Figure 1. Data were abstracted from each chart, including serum lab values, chylous effusion variables, and growth parameters. The collected data were stored and managed electronically using REDCap hosted by Tulane University. Thirty percent of data were cross-checked by another research staff member to ensure data accuracy and integrity. The data were analysed via SPSS.

Figure 1.

Participant flow chart.

Growth parameters collected included height, weight, and Centers for Disease Control and Prevention (CDC) growth percentiles at the following time points: 1) admission, 2) procedure, and 3) discharge. Total volume of blood products transfused, days on total parenteral nutrition, and doses of plasmalyte, received during admission were also collected as covariates to subjects’ weight measurements. The markers of nutrition collected were 1) serum total protein and 2) serum albumin. The chylothorax parameters collected included: 1) chylous triglyceride levels and 2) maximum daily chylous output in millilitres per kilogram per day. Additional health outcomes of interest were collected including post-operative complications, days on ventilation, length of stay, and mortality. Mortality was defined as within one year of follow-up.

Among cases, Shapiro–Wilk test confirmed normality of growth and nutrition parameters including: growth velocity, weight change, weight-for-length z-score change, serum protein, and serum albumin. Weight-for-length Z-scores were calculated based on the World Health Organization (WHO) calculator for infants < 24 months of age only. Student’s t-test was used to compare the aforementioned growth and nutrition parameters between cases and controls. Two-sample t-test was also used to compare the volume of blood products received, days on total parenteral nutrition, and doses of plasmalyte between cases and controls. Linear regression was used to ascertain the association between maximum chylous output with both growth velocity and weight change. Mann–Whitney U-Test was conducted to compare chylous triglyceride levels in cases with no weight gain or weight loss compared to subjects with weight gain over the hospital course.

Discussion

With nearly one-fourth of infants with univentricular heart defects satisfying failure to thrive criteria within the first year of life, ^{Reference Medoff-Cooper, Naim, Torowicz and Mott17} a clear understanding of the implications of CHD and its complications on early growth and nutrition is imperative. Our study is the first to develop a more comprehensive picture of growth and nutrition markers in children with and without chylothorax development post-CHD surgery. Contrary to previous studies, weight measurements and Z-scores were collected and analysed at multiple timepoints allowing more accurate and reliable representations of growth than discrete measurements. Z-scores were incorporated into our analysis to describe a subject’s growth relative to the population mean using the standard deviation of the population of interest, allowing for more accurate ascertainment of growth over the study period. ^{Reference George, Jagannath, Joshi and Jagadeesh18} The fluid covariate data revealed a greater degree of fluid resuscitation in the control group, which may indicate the chylothorax cases were fluid restricted. Differences in mean transfused volumes are worth considering when using weight as a marker for nutrition. The significant difference in weight gain between chylothorax and control cases during enteral feeds post total parenteral nutrition highlights how vital it is to optimise nutrition in chylothorax patients post CHD surgery, since the data indicate they are at greater risk of inadequate weight gain. Accurate representation of growth is crucial to determine adequate caloric intake in CHD infants with increased risk for failure to thrive. In this already vulnerable population, failure to thrive adds to the risk of long-term cognitive delay and poor neurodevelopmental outcomes. ^{Reference Mangili, Garzoli and Sadou5,Reference Brief, Guimber and Baudelet19}

Serum albumin is routinely measured during standard blood work, and the increased half-life reflects a dietary change over a three-week duration. Albumin is often considered a less reliable marker of nutrition than prealbumin because of its prolonged half-life and variability in patient with ascites, renal protein loss, liver disease, and inflammatory states. ^{Reference Helms, Dickerson, Ebbert, Christensen and Herrod20,Reference Ford, Jennings and Andrassy21} However, prealbumin levels drop rapidly in malnourished patients (half-life of two days) and normalise within ten days of adequate nutrition. ^{Reference Ingenbleek and Young22,Reference Elhasid, Laor, Lischinsky, Postovsky and Weyl Ben Arush23} Therefore, prealbumin measurements can potentially indicate whether a recent change in intervention or pathology may influence nutritional status and should be considered in lieu of albumin as a marker of nutritional status.

Children with CHD often have increased resting energy expenditure due to increased cardiac and respiratory effort. Inadequate energy intake from anorexia, early satiety, and feeding intolerance issues are also significant comorbid concerns. ^{Reference Varan, Tokel and Yilmaz24} Current protein guidelines exist for infants with CHD, but they vary broadly and fail to consider the unique needs of infants with postoperative chylothorax. ^{Reference DiLauro, Russell, McCrindle, Tomlinson, Unger and O’Connor25,Reference Floh, Slicker and Schwartz26} Serum total protein is a routine and affordable marker that can gauge a patient’s response to and tolerance of nutritional supplementation. It is worth noting that total protein synthesis is inhibited by proinflammatory states, which in and of themselves contribute towards malnutrition. These protein markers may potentially reflect the extent of inflammation in patients more so than the degree of malnutrition from a depletion in protein energy stores. ^{Reference Keller27}

Prolonged hospital length of stay and increased mortality are two potential consequences of acquired chylothorax post CHD surgery, which our data support. A 110-centre study determined a 1.4% to 2.9% estimated morality risk associated with STAT Category 2, and a 3.0% to 6.8% risk associated with STAT Category 3. With an average STAT score of 2.53, the morality rates seen in our patient population far surpass this study’s estimated risk. ^{Reference Jacobs, Jacobs and Thibault28} Previous studies found acquired chylothorax to be associated with longer length of stay, increased adjusted risk for in-hospital mortality and higher cost, regardless of procedure complexity. ^{Reference Mery, Moffett and Khan29,Reference Martin, McBrien, Marchak, Atallah, Al Aklabi and Mackie30} Our data support this finding. Extended length of stay have been shown to increase rates of nosocomial infections and long-term morbidity and mortality between 15 and 40%, which should be taken into considered in clinical practice and future studies. ^{Reference Ping Kirk, Sng, Zhang, Ming Wong, Puthucheary and Lee31–Reference Seo, Park and Yun34}

Our most notable limitation was our small sample size (n = 38). Increased sample size can be achieved for this rare complication via inter-institutional collaboration and multisite trials. Despite this, study inclusion criteria were developed to avoid convenience sampling. While case–control matching was manually performed, the patients were selected solely by closest age and shared procedure type. Greater length of stay noted in chylothorax cases may affect weight at the time of discharge and contribute to greater variability in collected growth metrics. Additional nutritional markers, including serum triglycerides, prealbumin, and fat-soluble vitamin levels, were not routinely collected at our institution and provide additional opportunities for future studies.

It would be reasonable to presume that damage to the thoracic duct and subsequent chylothorax would cause a drop in serum triglyceride levels, specifically long-chain triglycerides. ^{Reference Büttiker, Fanconi and Burger35} This is of potential significance for infants and children since triglycerides are needed for energy reserves, cholesterol synthesis, and because long-chain triglycerides are considered essential for maturation of the developing brain, retina, and other organs in newborn infants. ^{Reference Jasani, Simmer, Patole and Rao36} Future studies should seek to evaluate the relationship between chylothorax development and serum triglycerides. Finally, only six cases of chylothorax were considered high-volume chylothorax. While low output effusion is generally treated conservatively, high output effusions typically require more aggressive treatment. ^{Reference Yeh, Brown and Kellogg37} A larger sample size may have allowed stratification of low- and high-output cases.

Conclusion

Compared to controls, cases of chylothorax did not have statistically significant markers of growth or nutrition. For cases of chylothorax, maximum chylous output and chylous triglyceride level were not associated with markers of growth and nutrition in CHD patients with chylothorax. The increased length of stay and mortality seen in the chylothorax cases highlights the harmful impact this condition can have on CHD patients. The challenge of tailoring treatment for possible nutritional losses remains due to the paucity of comprehensive data on this topic.

Future studies should employ a multicentre, prospective study to facilitate an increase in sample size and power. We recommend that clinicians routinely measure key nutritional markers in serum and chylous output and record reliable growth metrics, since they are both relatively low-cost and lost-risk assessments that may improve medical management of these patients. Investigation is warranted on treatment methods (i.e., median chain triglycerides, total parenteral feeds, and surgery) and their influence on nutrition and growth, as well as morbidity and mortality. Multidisciplinary research and multicentre studies are needed to ascertain the clinical impact chylothorax has on the nutritional status of infants and children to better guide clinical practice.