The first-generation test, TDx FLM, was developed by Russell (39) at Abbott Laboratories, Inc., in collaboration with Tait et al. (40) as an attempt to improve on the original work of Shinitzky and colleagues (41). The original Shinitzky method used diphenyl-hexatriene (DPH) and an ultraviolet fluorescence spectrophometer. The results were reported in polarization units. Because DPH is very unstable and the instrument was difficult to maintain, this method never achieved widespread use. Tait's method uses a laboratory instrument common to most US laboratories, the Abbott TDx, and a fluorescent dye, NBD-PC. Russell's assay, TDx FLM, uses a more stable fluorescent dye, PC-16. In the early 1990s, Abbott Laboratories modified the first-generation test to produce the second-generation TDx FLM II.

Method Description

The TDx FLM II method uses uncentrifuged AF that is passed through a glass filter. While the TDx FLM method used 0.5 mL of filtered AF, the TDx FLM II method uses only 0.25 mL. The filtration step, however, usually requires 1 mL. The method is calibrated with six solutions containing differing ratios of surfactant to albumin. The results are reported in units of mg of surfactant to g of albumin. The analytical measurement range is 0 to 160 mg/g. The method has excellent precision of 4% at 20 mg/g and 5% at 160 mg/g. The transition from TDx FLM to TDx FLM II apparently lowered the results, although no direct comparison can be found in the published or manufacturer literature.

The NBD-PC method, known simply as FPol, has a very strong nonlinear correlation to PC-16. The FPol results are reported in mPol units and vary from 160 mPol to 350 mPol. Unlike the TDx FLM II, the values decrease with increasing maturity. The FPol method uses mild centrifugation at 400g for 2 min. The strong correlation (r = 0. 95) permits outcome studies by FPol to be converted into the units of the TDx FLM (42). From unpublished work in my research laboratory, the equation for converting from FPol to TDx FLM II results is

TDx FLM II (in mg/g) = -139.48 ln(FPol in mPol) + 817.1

The FPol decision limits are mature, less than 260; transitional, 260 to 290; and immature, greater than 290 mPol. Note that 260 mPol is equivalent to 41.5 mg/g and that 290 mPol is equivalent to 26.2 mg/g.

Preanalytical Variables

Although uncentrifuged AF should be used, laboratories can resuspend fluids that have been accidentally centrifuged (43). Either mild vortexing or thorough mixing for 5 min on a tube rocker suffices.

Blood contamination tends to lower results greater than 55 mg/g and increase results less than 39 mg/g (43,44). Although blood contamination changes the results, an immature result will not erroneously be classified as mature; 5% blood contamination increased an immature result of 18 mg/g to 32 mg/g, whereas an indeterminate result of 48 mg/g was unchanged (43).

A 2002 study reported that TDx FLM of AF were unchanged following 16 h at room temperature and 24 h at 4°C (43). Freezing at -20°C for up to 11/2yr introduced an average negative bias, but, more importantly, showed a large random error. Results were ±30% of their baseline values. This is a striking contrast to the stability of FPol in samples stored at -20°C (45).

The use of specimens collected vaginally has been shown to produce results that are on average 35% lower than those collected by amniocentesis (46). Thus, a mature result from a vaginal pool specimen can be trusted, but an immature result may be falsely lowered.


Abbott Laboratories recommends three interpretation categories for TDx FLM II: immature (<39 mg/g), intermediate (40-54 mg/g), and mature (>55 mg/g). The maturity decision limit for the first-generation TDx FLM was higher; 70 mg/g (47). Several studies of the first-generation test suggested that the upper decision limit could be lowered without reducing the sensitivity for RDS (48-50). The second-generation TDx FLM II

also appears to have a conservative upper decision limit (51-55). Therefore, contrary to the manufacturer's recommendation, an upper decision limit of greater than or equal to 45 mg/g is recommended because it will improve specificity to about 85% without compromising high sensitivity. Although the exact confidence interval for the sensitivity at this cutoff is difficult to estimate because of the small aggregate number of RDS cases reported (<100), it is probably 95-100%. The 45 mg/g decision limit agrees well with clinical outcome studies of FPol (56-59) that show that 260 mPol yields a sensitivity of 95% and specificity of approx 65%.

Several investigators have recommended that TDx FLM II be interpreted after gestational age stratification (60-62). Two recent editorials have also encouraged this practice (63,64). Instead of using differing decision limits at each gestational age, the authors advocate reporting the risk of RDS for the TDx FLM II result in combination with the gestational age. The logistic model recommended by one editorial (64) was derived from the first-generation TDx FLM, making its use with TDx FLM II inadvisable (65). Because these risks are essentially predictive values, they will vary with the tested population. Because the studies were conducted at tertiary hospitals that handle high-risk obstetrical cases, use of these models is best restricted to those settings. If the models are applied to hospitals that care for low-risk obstetrical populations, the risks reported should be disclaimed as maximum risks.

Almost half of all twins are delivered preterm. If complications other than preterm are excluded, twins appear to have the same incidence of RDS as do gestational age-matched singletons (66). Yet after 31 wk gestation, the average TDx FLM II results from twin pregnancies are 22 mg/g higher than those from singleton pregnancies (67). The ability of TDx FLM II to predict RDS in twins may be different than its predictability in singletons. If testing prior to 32 wk gestation, sampling both sacs is recommended (68).

Several studies have shown that the TDx FLM II is reliable when used in diabetic pregnancies without changing the reference values (69-71).

0 0

Post a comment