Our objective was to directly compare the diagnostic ... mature result was 98%, and predictive value of an ... ferent analytic techniques with the former measuring rotation .... vals (CIs) were calculated using Prism 4.0 software (GraphPad.
Clinical Chemistry / LBC AND TDX-FLM II FOR PREDICTING RDS
A Direct Comparison Between Lamellar Body Counts and Fluorescent Polarization Methods for Predicting Respiratory Distress Syndrome Shannon Haymond, PhD,1* Veronica I. Luzzi, PhD,2 Curtis A. Parvin,1 and Ann M. Gronowski, PhD1 Key Words: Lamellar body counts; Fluorescent polarization; Fetal lung maturity DOI: 10.1309/8VXN5EM5L3831AT2
Abstract Our objective was to directly compare the diagnostic usefulness of lamellar body counting (LBC) and the TDx-FLM II assay (Abbott Laboratories, Abbott Park, IL) for predicting respiratory distress syndrome (RDS). This was a 5-year, retrospective, cohort study. A diagnosis of RDS was given to infants who received surfactant treatment and/or required ventilator support and/or continuous positive airway pressure for more than 24 hours. There were 172 infants without RDS and 12 with RDS included in the study. By using a TDx-FLM II cutoff of 55 mg/g or more for maturity, the sensitivity was 83%, specificity was 65%, predictive value of a mature result was 98%, and predictive value of an immature result was 14%. These results were similar to LBC using a cutoff of 50,000/µL or more with sensitivity of 92%, a specificity of 60%, a predictive value of a mature result of 99%, and a predictive value of an immature result of 14%. The LBC and TDx-FLM II methods have similar clinical usefulness.
894 894
Am J Clin Pathol 2006;126:894-899 DOI: 10.1309/8VXN5EM5L3831AT2
Respiratory distress syndrome (RDS) is a life-threatening illness that occurs in infants delivered before completion of fetal lung maturation. RDS is caused by insufficient production of pulmonary surfactant by type II pneumocytes in the developing lung. When surfactant is absent, small alveoli collapse and larger alveoli are overinflated, resulting in a stiff, noncompliant lung. In 2003, the infant death rate due to RDS in the United States was 20 per 100,000 live births.1 Because fetal lung liquids contribute to amniotic fluid, the amount of surfactant present in fetal lungs can be estimated by measuring the amount of surfactants in amniotic fluid. The fluorescence polarization method for estimating surfactants in amniotic fluid is based on the competitive binding of a fluorescent probe to albumin and surfactant. When the probe is bound to albumin, polarization is high. When the probe is bound to surfactant, polarization is low.2 The degree of net polarization is correlated with fetal lung maturity. The automated commercial fluorescence polarization assay (TDxFLM II, Abbott Laboratories, Abbott Park, IL) is the most common quantitative method for assessing fetal lung maturity.3 Several studies have examined the clinical usefulness of the TDx-FLM II assay and have shown that it has excellent sensitivity4-7 that ranges from 95.7% to 100% and specificity between 70% and 72% using the manufacturer’s recommended cutoff of 55 mg/g. Lamellar bodies represent the storage form of pulmonary surfactant within type II pneumocytes. The similarity of lamellar body size to platelet size permits the use of a standard hematologic counter to quantify lamellar body concentration in amniotic fluid. This technique, known as lamellar body count (LBC), is used to estimate surfactant production in utero and, thus, predict the degree of fetal lung maturity. The ability to © American Society for Clinical Pathology
Clinical Chemistry / ORIGINAL ARTICLE
perform an LBC on a hematology analyzer makes it widely available, inexpensive, rapid, and easy to perform and requires a low sample volume. The majority of published studies have used the Beckman Coulter (Brea, CA) hematology analyzer. Research has demonstrated that there are differences in LBCs when different analyzers are used, and, as a result, analyzerspecific reference intervals should be established.8 Many outcomes studies9-19 and 1 meta analysis20 examined the clinical usefulness of the LBC. These studies demonstrated that the performance of LBC is equal to or better than the well-established L/S ratio. The reported sensitivity of LBC ranges from 83% to 100% and specificity ranges from 54% to 89% depending on the method and cutoff used.2 In 1996, the American College of Obstetrics and Gynecology recommended a “cascade” or “sequential” approach to testing for fetal lung maturity.21 This was recommended because a mature result on any of the commonly used tests for fetal lung maturity is strongly predictive of the absence of RDS, and little additional information is to be gained by performance of additional assays. It is recommended that a rapid method be performed first, and, if the result is clearly immature or clearly mature, no further testing is required. Fluorescence polarization and LBC are the most rapid quantitative methods for assessing fetal lung maturity. The fluorescence polarization and LBC methods are completely different analytic techniques with the former measuring rotation of a dye when associated with phospholipids and the latter quantifying the number of multilamellar phospholipid packages. Despite these differences in analytic method, there has not been a careful systematic comparison between the clinical usefulness of the TDx-FLM II and LBC using the Coulter method. The present study is the largest of its kind, with outcomes data, to directly compare the clinical usefulness of LBC using the most widely used hematology analyzer (Beckman Coulter) with the TDx-FLM II assay.
Materials and Methods Samples This study was a 6-year (February 1998 to March 2004) retrospective study using 2 sources of samples: (1) all patients who had physician-ordered TDx-FLM II and LBC testing ordered on the same amniotic fluid (n = 87; 6 RDS and 81 nonRDS) and (2) randomly selected, leftover amniotic fluid samples sent to the laboratory for physician-ordered TDx-FLM II testing. Samples were chosen with no knowledge of TDx-FLM II result or patient outcome. An LBC was determined for each specimen within 7 days of collection. No attempt was made to seek cases of RDS (n = 97; 6 RDS and 91 non-RDS).
Study inclusion criteria required that delivery occurred within 72 hours of the amniocentesis and that medical charts were available for mother and infant. If steroids were administered after amniotic fluid collection and the infant was not affected with RDS, the sample was excluded. In these cases, it was impossible to know whether the steroids impacted the respiratory status of the infant. If steroids were administered after amniotic fluid collection and the infant was affected with RDS, the sample was included because steroid administration did not impact the respiratory status of the infant. Although steroids may have tempered the severity of the RDS, data were analyzed only for the presence or absence of RDS. All multiple gestations were excluded. The diagnosis of clinically significant RDS was given to infants who were treated with surfactant and/or were placed on a mechanical ventilator for more than 24 hours and/or required continuous positive airway pressure for more than 24 hours. Institutional review board approval was obtained for the study. Analytic Methods Amniotic fluid samples were collected either by transabdominal amniocentesis or from the free flowing vaginal pool if the membranes were ruptured. The manufacturer of the TDx-FLM II assay considers both specimens acceptable for analysis. Samples containing visible blood, bilirubin, or meconium were rejected. The TDx-FLM II analysis was conducted according to the manufacturer’s instructions. The upper and lower limits of detection for the TDx-FLM II assay were 160 and 10 mg/g, respectively. The cutoff used for indicating maturity by the TDx-FLM II was 55 mg/g or more. The LBC was performed using the platelet channel of a Coulter Gen-S (Beckman Coulter) hematology analyzer. The upper and lower limits of detection for the Coulter Gen-S instrument on the platelet channel were 3 million/µL and 2,000/µL, respectively. The mature cutoff used for the LBC assay was 50,000/µL or more.22 Samples were stored between TDxFLM II and LBC analysis at 4°C for fewer than 7 days and were not centrifuged before analysis. LBCs have been shown to be stable at 4°C for up to 10 days.23 When available, lecithin/sphingomyelin (L/S) and phosphatidyl glycerol (PG) data were included for comparison. The L/S and PG analyses were performed by thin-layer chromatography at Quest Diagnostics (St Louis, MO, method24-26). Diagnostic cutoffs were as follows for L/S: immature, less than 1.0; premature, 1.0 to 1.5; intermediate, 1.5 to 1.9; caution mature, 2.0 to 2.5; and mature, more than 2.5. The PG/S analysis was interpreted as follows: immature, less than 0.3; and mature, 0.3 or more. Statistical Analysis Sensitivity is the probability of observing a result less than the cutoff when clinically significant RDS develops, and Am J Clin Pathol 2006;126:894-899
© American Society for Clinical Pathology 895
DOI: 10.1309/8VXN5EM5L3831AT2
895 895
Haymond et al / LBC AND TDX-FLM II FOR PREDICTING RDS
specificity is the probability of observing a result greater than or equal to the cutoff when clinically significant RDS does not develop. The predictive value of a mature result (or negative predictive value) is the probability that clinically significant RDS does not develop when a result greater than the cutoff is obtained. The predictive value of an immature result (or positive predictive value) is the probability that clinically significant RDS will develop when a result less than the cutoff is obtained. Receiver operating characteristic (ROC) curves were drawn, and the areas under the ROC curves and 95% confidence intervals (CIs) were calculated using Prism 4.0 software (GraphPad Software, San Diego, CA). Areas under ROC curves were statistically compared using the method of DeLong et al.27
Results During the study period, 407 amniotic fluid samples were available for inclusion. Of the initial 407 samples, 223 did not meet the study inclusion criteria. We excluded 69 samples because of insufficient patient information and 132 because delivery occurred more than 72 hours after the amniocentesis procedure. Samples for 7 infants with congenital anomalies and 12 cases of multiple births were excluded. Steroid administration between the time of sample collection and delivery resulted in the exclusion of 3 cases. Thus, 184 samples met the inclusion criteria. A total of 184 cases were included in the study, of which 172 were categorized as unaffected, non-RDS and 12 as affected, RDS. The 6.5% prevalence of RDS in this study is similar to that reported in other studies.4,7,9,15 The characteristics of the 12 patients with RDS are shown in ❚Table 1❚. The 12 RDS cases had LBCs ranging from
5,000/µL to 52,000/µL, with 1 patient with an LBC greater than the cutoff for maturity of 50,000/µL or more. The TDxFLM II ratios for the 12 RDS cases ranged from 12.0 to 108.4 mg/g, with 2 patients having TDx-FLM II ratios greater than the cutoff for maturity of 55 mg/g or more (70.3 and 108.4 mg/g). No RDS case had a mature result for PG/S (data available for only 6 cases). Of the 6 cases with a measured L/S, 3 RDS cases had L/S values in the caution mature range, and 3 cases were in the intermediate range. In 2 cases, infants were delivered with RDS with LBC and/or TDx-FLM II results greater than the cutoff for maturity and were considered to have clinically relevant RDS as dictated by the study criteria; thus, there was no reason for exclusion. These 2 cases are described here. In case 11, a premature boy was delivered by elective cesarean section at a 36-1/7 week estimated gestational age to a 31-year-old gravida 2, para 2 mother whose pregnancy was complicated by poorly controlled type 1 diabetes mellitus. Following a tachypneic episode in the delivery room, the infant received continuous positive airway pressure for 1 minute, and improvement in color was noted. He then required blow-by oxygen at 20 minutes of life. He was transferred to the special care nursery for further management after 3 unsuccessful attempts to wean him to room air. In the special care nursery, he was given oxygen under a 30% oxy-hood and then received oxygen at 0.3 L/min per nasal cannula. The RDS resolved, as he was weaned to room air without problems on the third day of life. In case 12, a premature boy was delivered by repeated cesarean section at a 36-2/7 week estimated gestational age to a 39-year-old gravida 5, para 5 mother. In this infant, mild to moderate respiratory distress developed in the delivery room, and he was transferred to the special care nursery
❚Table 1❚ TDx-FLM II and Lamellar Body Count Results for Infants With RDS Case No. 1 2 3 4 5 6 7 8 9 10 11 12
Gestational Age (wk)
TDx-FLM II Assay (mg/g)*
Lamellar Body Count (/µL)†
L/S Ratio‡
PG/S Ratio§
34-4/7 34-6/7 34-1/7 36 36-4/7 36-1/7 36-4/7 32 36-4/7 35-4/7 36-1/7 36-2/7
12.0 22.9 27.3 28.8 31.1 32 34.7 35.2 35.6 36.8 70.3 108.4
5,000 6,000 23,000 17,000 31,000 22,000 30,000 19,000 22,000 15,000 43,000 52,000
ND 1.7 1.8 ND 2.3 ND 1.6 2.4 2.2 ND ND ND
ND
