Krishnan, Usha*; Mitchell, John D.†; Tobias, Vivian‡; Day, Andrew S.*§; Bohane, Timothy D.*
Aspiration of refluxed gastric material results in a variety of respiratory problems, including recurrent aspiration pneumonia, asthma, chronic cough, stridor, and apnea (1). Although simultaneous esophageal and tracheal pH probe monitoring (2,3) demonstrates a temporal relation between gastroesophageal reflux (GER) and respiratory problems, current diagnostic techniques do not consistently identify or rule out aspiration. Both technetium-labeled milk scan and barium swallow assess events over a short time frame, markedly reducing the diagnostic reliability of negative studies in the diagnosis of GER aspiration, which is probably of a sporadic and intermittent nature (4–6).
The identification of fat-laden macrophages (FLM) in tracheal aspirates has been used for many years as a diagnostic test for pulmonary aspiration (7–14). This test relies on the identification of fat in the respiratory system after ingestion by macrophages (15). The clinical utility of this assay has been questioned because of an unacceptable incidence of false-positive results and its failure to detect recurrent aspiration of material free of fat (12) or to distinguish aspiration of gastric contents from direct aspiration of material from the oropharynx. The aim of this study was to reevaluate the role of this assay in diagnosing reflux aspiration.
Ninety-eight infants and children undergoing general endotracheal anesthesia at Sydney Children's Hospital were recruited in this prospective and case-controlled study during a 12-month period at the Sydney Children's Hospital. Of the 98 children, 84 were undergoing elective or emergency gastrointestinal endoscopies. Of the 84 children, 64 were being reassessed to clarify the status of clinically significant GER, and the remaining 20 had a variety of other reasons (inflammatory bowel disease, n = 12; bleeding per rectum, n = 4; foreign body removal, n = 2; and celiac disease, n = 2). The remaining 14 were undergoing anesthesia for elective cardiac surgery (n = 8), ear/nose/throat surgery (n = 4), orthopedic surgery (n = 1), and general surgical procedures (n = 1).
All children were allocated before anesthesia to one of four clinical groups depending on presence or absence of clinically significant reflux (GER+ or GER−, respectively) or chronic respiratory symptoms (RS+ or RS−, respectively). A questionnaire modified from Orenstein's validated Infant Gastroesophageal Reflux Questionnaire (13) was used to determine the presence of clinically significant reflux. Children were categorized as having “chronic respiratory symptoms” based on the presence of persistent cough and/or asthma for more than 3 months or recurrent episodes of “bronchitis” or pneumonia. Two children with difficult-to-treat cystic fibrosis lung disease were included in this group as they had both GER and respiratory symptoms. A history of chronic respiratory symptoms was obtained in 45 of the 98 children. There were 37 children in the GER+RS+ group, 27 in the GER+RS− group, 26 in the GER−RS− group, and 8 in the GER−RS+ group. Details of previous investigations (chest radiograph, barium meal, pH probe, and/or endoscopy) were completed with the help of the parent or guardian and patient records (Table 1). The timing of investigations before the date of collection of tracheal aspirate samples varied from 1 week to, on one occasion, 2 years (this was a pH probe study in a neurologically delayed child with symptoms of reflux). However, the median time between the timing of investigations and collection of tracheal aspirate was 3 months.
Reflux esophagitis was judged to be present if the basal layer of cells was greater than 25% of epithelial thickness, papillary height was greater than 50% of epithelial thickness, and eosinophils or ulceration was present.
The Research and Ethics Committee of the South Eastern Sydney Area Health Service approved the study, and informed consent was obtained from a child's parent or guardian before inclusion in the study.
Tracheal aspirates obtained from the same group of patients were also assayed for gastric pepsin as a new marker for reflux aspiration (data to be published as a separate study).
Collection of Tracheal Aspirate
Tracheal aspirates were obtained from each patient after endotracheal intubation. Time between commencement of induction and intubation varied between 3 and 15 minutes, usually dependent on difficulty in obtaining venous access. After intubation and before the start of the procedure or operation, an appropriately sized “Y” suction catheter, with one limb connected to a small sterile mucous trap, was passed through the endotracheal tube into the trachea proximal to the carina. Two milliliters of normal saline was instilled through the catheter and then rapidly suctioned into the mucous trap. The aspirate sample was immediately frozen and stored at −80°C till assayed for pepsin. In distinction to a standard bronchoalveolar lavage, the amount of saline instilled for collection of the tracheal aspirate sample was smaller and was not instilled under pressure. Consecutively, the sample collected was considered to be more representative of tracheal fluid rather than respiratory secretions from terminal and respiratory bronchioles, which contain more cellular material.
Histopathologic Examination of the Tracheal Aspirates
Tracheal aspirates were cytocentrifuged, and four slides were prepared from each sample. One of the slides was stained with Papanicolou stain, and the remaining three were stained by a modified Oil Red O stain.
Cellular Lipid Evaluation
To determine the amount of intracellular lipid, the lipid cellular index was computed based on the study by Corwin and Irwin (8). The amount of lipid in macrophages was graded from 0 to 4 as follows: grade 0 = absence of any intracellular fat droplets; grade 1 = presence of one to a few intracellular fat droplets; grade 2 = presence of many distinct intracellular fat droplets; grade 3 = presence of many confluent intracellular fat droplets; grade 4 = presence of many confluent intracellular fat droplets, completely opacifying the cytoplasm and obscuring the nucleus.
For each tracheal aspirate specimen, three slides were prepared using a modified Oil Red O stain. Grading of 100 macrophages was spread evenly over three slides, and the sum of the grade for each macrophage constituted the lipid cellular index (LCI). Because the maximal grade for each macrophage is 4, the maximum possible LCI is 400. An LCI of 100 was used as a cutoff point between positive and negative (8).
The χ2 test, Student t test, and Fisher exact test were used to evaluate the results. Probability statistics were calculated using SPSS statistical software (SPSS Inc., Chicago, IL, U.S.A.) and were used to describe the testing characteristics of the lipid cellular index in terms of sensitivity, specificity, false-positive rate, false-negative rate, positive predictive value, negative predictive value, and overall efficiency.
The mean age ± SEM of the four groups was 6 ± 0.9 years in the GER+RS+ group, 7.5 ± 1.8 years in the GER+RS− group, 6.7 ± 0.10 years in the GER−RS+ group, and 6.11 ± 0.8 years in the GER−RS− group (P = NS). The male:female ratio in the four groups was similar except for a male preponderance in the group with both reflux and respiratory symptoms (28 males, 9 females). Thirty-seven children in the GER+ group had chronic respiratory symptoms of asthma (n = 20), chronic cough (n = 7), or recurrent chest infections (n = 10). Eight children in the GER− group had chronic respiratory symptoms of either asthma (n = 5) or chronic cough (n = 3). Spectrum bias was avoided in the GER+ patients by including consecutive endoscopies. The histologic finding of esophagitis in 42% of the GER+ patients suggests the majority did not have severe reflux disease. Presence or absence of respiratory symptoms was not a criterion for selection into the GER+ group, thereby avoiding selection bias. Observer bias was also avoided by blinding the histopathologist blind to the grouping of patients. Details of previous investigations (chest radiograph, barium meal, pH probe, and/or endoscopy) completed with the help of the parent or guardian and patient records, are shown in Table 1. In the GER+ group, 27 of the 64 children had 24-hour pH probe studies performed. All had reflux indices outside the reference range (16). Thirty-nine had a barium meal, and reflux was seen in 25. However, no radiologic evidence of reflux aspiration was observed on barium study in the 19 children with clinically significant reflux and respiratory symptoms. Esophagitis was proven histologically in 25 children.
The mean LCI in the four groups did not differ: 107 ± 15 in the GER+RS+ group, 76 ± 15 in the GER+RS− group, 94 ± 27 in the GER−RS+ group, and 89 ± 14 in the GER−RS− group (P = NS). Furthermore, there was no difference in the frequency of children with a mean LCI of 100 or greater between the four groups: 16/37 in the GER+RS+ group, 8/27 in the GER+RS− group, 4/8 in the GER−RS+ group, and 10/26 in the GER−RS− group (P = NS). These results were further analyzed to calculate the diagnostic sensitivity (38%), specificity (59%), positive predictive value (63%), negative predictive value (33%), false-positive rate (41%), false-negative rate (63%), and overall efficiency (45%) of the assay. It is of interest that of 60 children with an LCI less than 100, 35% belonged to the GER+RS+ subgroup, the group considered most likely to be at risk of reflux aspiration. Conversely, of 31 children with an LCI of 100 or greater, 32.2% belonged to the subgroup that was negative for both GER and respiratory symptoms, and therefore considered very unlikely to be aspirating.
Interestingly, when an LCI of 75 or greater was used as the cutoff, the sensitivity of the test increased marginally (41%), but the specificity further decreased (47%). Similarly, an LCI of 125 or greater increased the specificity of the assay (68%), but the sensitivity further decreased (33%) and the false-negative rate increased (67%).
Table 2 compares findings of the investigations for assessing GER and/or aspiration with the results of the assay for FLM. There were no significant differences in either the prevalence or number of FLM detected between the children either with or without histologic esophagitis or GER seen during a barium study. The frequency and number of FLM was not significantly higher, even when children in either of the preceding groups also had respiratory symptoms. Of 64 children in the GER+ group, 29 had esophagitis on histology. Of these 29 children, 17 had mild, 8 moderate, and 4 severe esophagitis histologically (including two with ulcerative esophagitis). The number of children with an LCI of 100 or greater was similar in those with mild (8 of 17) or moderate esophagitis (4 of 8) (P = NS). However, 3 of 4 patients with severe esophagitis had an LCI of 100 or greater. The two children with ulceration also had chronic respiratory symptoms, and both of their LCIs were 100 or greater. Nine of the 35 children for whom reports of a chest radiograph were available had changes consistent with or suggestive of pulmonary aspiration. All nine had symptomatic reflux and recurrent lower respiratory infections, and all were positive for FLM. However, two children with poorly controlled cystic fibrosis–related lung disease and excessive reflux on a pH probe study tested negative for FLM.
The aspiration of gastric contents causing pneumonia is often equated with exogenous lipoid pneumonia, in which lipid-laden macrophages are an integral part of the histopathologic picture. However, there is still a question regarding the mechanism by which microaspiration of gastric content results in accumulation of fat within the alveolar macrophage (15). Indeed, the rate of macrophage intracellular metabolism of ingested lipid and the timing of macrophage egress from the lung is unknown (1,17). Rapid clearance of fat by the macrophage would restrict the effective time frame of the lipid index. As a result of these limitations, the diagnosis of reflux microaspiration may be delayed.
Clinical studies have suggested that the occurrence of FLM in bronchoalveolar lavage fluid correlates well with the clinical diagnosis of GER complicated by aspiration pneumonia (8–11). However, a 1985 report by Staugas et al. (12), examining tracheal aspirates of children with and without proven GER, determined that 42% of children in the control group with no history of either lower respiratory tract disease or GER were positive for FLM. In this study, no attempt was made to quantify FLM. A later study by Moran et al. (7) examined the tracheal aspirates obtained from newborn infants requiring mechanical ventilation and concluded that FLM could be present in neonates without aspiration, and that their presence was a nonspecific finding. More recently, Adam et al. (14) concluded that the FLM index was a useful indicator only for aspiration, which caused radiologic lung disease in adults, and that this index lacked specificity.
The aim of the current study was to quantify the presence of FLM in tracheal aspirates of children with and without a history of GER and/or respiratory symptoms to reevaluate the role of FLM as a marker of reflux aspiration. Our results suggest that FLM do not occur more frequently in the tracheal aspirates of children with symptomatic reflux (GER+) than in those without such symptoms (GER−), and that FLM are not even found more commonly in children with coexistent GER and respiratory symptoms (GER+RS+) than in children without either of these symptom complexes (GER−RS−). Furthermore, there was no significant difference in the mean LCIs of the groups of the children with and without symptomatic reflux. These results were supported by a lack of sensitivity (38%) and specificity (59%) of the FLM assay.
One neurologically developmentally delayed child in this study with a history of recurrent aspiration pneumonia and who was receiving feedings by the mouth, tested significantly positive for FLM and subsequently improved (did not have any further episodes of pneumonia) after the placement of a feeding gastrostomy for bolus feeds without fundoplication. This suggests that FLM does not distinguish between reflux aspiration and direct aspiration caused by uncoordinated swallowing.
Two children with cystic fibrosis and poorly controlled respiratory disease had symptoms suggestive of GER and pathologic reflux on a 24-hour pH probe study, but neither was positive for FLM. One would have reasonably expected a high lipid index in such patients either because of reflux aspiration or from chronic airway inflammation resulting in leakage of endogenous lipids and decreased clearance of macrophages. The hypothesis that bacteria use lipids may provide an explanation for this phenomenon (9).
Esophageal pH probe monitoring, barium contrast studies, and endoscopic grading of esophagitis at best suggest the existence of a situation that may predispose to aspiration. Conclusions made regarding an association between GER and aspiration using these methods are assumptions, since an alternative mechanism of direct aspiration from the oropharynx secondary to neurologic or anatomic abnormalities may be the sole or major cause of the respiratory disease.
The fact that the prevalence and number of FLM were not significantly higher in children with abnormal parameters on a pH probe study, esophagitis on biopsy, or reflux seen during a barium study suggests that FLM is not a reliable marker of risk of reflux aspiration. This held true even of children in any of the preceding groups who also had respiratory symptoms (Table 2).
Other drawbacks in the assay for FLM are that it suffers from observer bias and lack of stability and reproducibility. It remains semiquantitative even after computation of an LCI. The assay must be performed on the same day as the collection of the aspirate as there is loss of fat from the cells. In addition, most of the cells in the aspirate degenerate with time and after thawing of frozen specimens. Even when slides are fixed and stained on the day of collection, the fat stains fade over a period of weeks, and restaining does not restore their original state. Because of these problems with storage, the assay is not reproducible. In the current study, there was no potential for observer bias in assessing and quantitating the lipid-laden macrophages and cellular lipid evaluation, as the sole pathologist was blinded to the clinical groupings of the subjects. The potential for observer error was limited by the fact that the pediatric histopathologist has more than 30 years' experience in this field.
The presence of FLM in the tracheal aspirates of children without either reflux or respiratory symptoms in this study remains unexplained. It may simply be that endogenous lipids are normally released during normal cell turnover. We are unable to comment on the source of the lipid within macrophages in these children, as there was no morphologic difference in the appearance of cellular lipid whether exogenous or endogenous in origin. The absence of any morphologic difference between exogenous and endogenous lipid in macrophages is consistent with the observation of Corwin and Irwin (8) but differs from that of other investigators (18–20). However, if the fat in the tracheal aspirate of children in the GER−RS− group is a result of microaspiration, it is conceivable that pulmonary microaspiration occurs in normal subjects (21) and that the symptoms of aspiration are simply proportional to the quantity aspirated.
In conclusion, we feel that the assay of FLM in tracheal aspirates is neither a specific nor a sensitive marker of reflux aspiration, nor does it elucidate the possible relation between GER and respiratory disease.
The authors thank Keith Kelly (Department of Anesthesia, Sydney Children's Hospital) and Associate Professor Alan Stark (Department of Statistics, University of New South Wales).
1. Guill MF. Respiratory manifestations of gastroesophageal reflux in children. J Asthma 1994; 32:173–89.
2. Donnely RJ, Berrisford RG, Jack CIA, et al. Simultaneous tracheal and oesophageal pH monitoring: investigating reflux associated asthma. Ann Thorac Surg 1993; 50:1026–34.
3. Jack CIA, Canerley PMA, Donnely RJ, et al. Simultaneous tracheal and oesophageal pH measurements in asthmatic patients with gastro-esophageal reflux. Thorax 1995; 50:201–4.
4. Balson BM, Kravitz KS, McGeady SJ. Diagnosis and treatment of gastroesophageal reflux in children and adolescents with severe asthma. Ann Allergy Asthma Immunol 1998; 81:159–64.
5. Ghaed N, Stein MR. Assessment of a technique for scintigraphic monitoring of pulmonary aspiration of gastric contents in asthmatics with gastroesophageal reflux. Ann Allergy 1979; 42:306–8.
6. Fawcett HD, Hayden CK, Adams JC, et al. How useful is gastroesophageal reflux scintigraphy in suspected childhood aspiration? Pediatr Radiol 1988; 18:311–3.
7. Moran JR, Block SM, Lyerly AD, et al. Lipid laden alveolar macrophages and lactose assay as markers of aspiration in neonates with lung disease. J Pediatr 1988; 112:643–5.
8. Corwin RW, Irwin RS. The lipid laden alveolar macrophages as a marker of aspiration in parenchymal lung disease. Am Rev Respir Dis 1985; 132:576–81.
9. Collins KA, Geisinger KR, Wagner PH, et al. The cytologic evaluation of lipid laden alveolar macrophages as an indicator of aspiration in young children. Arch Pathol Lab Med 1995; 119:229–31.
10. Nussbaum E, Maggi JC, Mathis R, et al. Association of lipid laden alveolar macrophages and gastro-oesophageal reflux in children. J Pediatr 1987; 110:190–4.
11. Colombo JL, Hallberg TK. Recurrent aspiration in children: Lipid laden alveolar macrophage quantitation. Pediatr Pulmonol 1987; 3:86–9.
12. Staugas R, Martin AJ, Binns G, et al. The significance of fat filled macrophages in the diagnosis of aspiration associated with gastro-esophageal reflux. Aust Paediatr J 1985; 21:275–7.
13. Orenstein SR, Shalaby TM, Cohn JF. Reflux symptoms in 100 normal infants: Diagnostic validity of the infant gastroesophageal reflux questionnaire. Clin Pediatr 1996; 35:607–14.
14. Adam R, Ruffin R, Campbell D. The value of lipid laden macrophage index in the assessment of aspiration pneumonitis. Aust NZ J Med 1997; 27:550–3.
15. Fisher M, Roggli V, Merten D et al. Coexisting endogenous lipid pneumonia, cholesterol granulomas, and pulmonary alveolar proteinosis in a pediatric population: A clinical, radiographic and pathologic correlation. Pediatr Pathol 1992; 12:365–83.
16. Vandeplas Y, Sacré-Smits L. Continuous 24 hour oesophageal pH monitoring in 285 asymptomatic infants 0–15 months old. J Paediatr Gastroenterol Nutr 1987; 6:220–4.
17. Bauer ML, Figueroa-Colon R, Georgeson K, et al. Chronic pulmonary aspiration in children. South Med J 1993; 86:789–95.
18. Genereaux GP. Lipids in lungs radiologic-pathologic correlation. J Can Assoc Radiol 1970; 21:2–15.
19. Sundberg RM, Krishna KE, Brown MJ. Evaluation of lipid pneumonia. Dis Chest 1959; 36:594–601.
20. Ribaudo CA, Grace WJ. Pulmonary aspiration. Am J Med 1971; 50:510–20.
21. Huxley EJ, Viroslav J, Gray WR, et al. Pharyngeal aspiration in normal adults and patients with depressed consciousness. Am J Med 1978; 64:564–8.
© 2002 Lippincott Williams & Wilkins, Inc.