Dr. Salwa Peracha sent several thought-provoking questions about PTs and PTTs, which are presented (at some length) on the next page:
We perform the prothrombin time (PT) and partial thromboplastin time (PTT) manually. We consider a PT that is prolonged greater than 3 seconds past the upper limit of the reference interval as abnormal and a PTT greater than 7 seconds above the upper PTT limit as abnormal. To elaborate, If the reference interval for PT is 11 to 14 seconds, then PT will be considered abnormal if it is 18 seconds. Likewise, if the reference range for PTT is 27 to 40 seconds, the PTT is considered abnormal if it is 48 seconds. Is this correct? What do you or others consider an abnormal PT and PTT?
Further, in the case of neonates up to 6 months of age, how do you report PT and PTT results? What is considered a prolonged PT or PTT? Is it possible to establish reference intervals with change in the lot of reagents?
Is it necessary to do a PT and PTT on 20 healthy individuals during the same day to establish reference intervals or can we do the testing as we get volunteers in 3 to 5 days?
Dr. Peracha, thank you for your questions.
As a generalization, I recommend you consider any PT or PTT result that is prolonged beyond the upper limit of the reference interval (reference range, normal range) to be abnormal. There is no statistical precedent for establishing an arbitrary interval beyond the upper limit that is still considered to be normal.
To equivocate, however, I draw your attention to our July, 2010 Quick Question about PTT mixing studies, where we learned that only 43% of respondents perform PTT mixing studies when the PTT is prolonged just past the upper limit, and that 15% perform mixing studies only when the PTT is “substantially” prolonged. Substantially, in the question, is undefined. Also, another 28% perform mixes only after collecting clinical information on the patient. In my commentary on the Quick Question results I referenced a comment by Dr. Dot Adcock that there was little to be gained by performing mixing studies when the result is just a second or two over the upper limit.
Consequently, the question of exactly how much past the upper limit the PTT has to be before doing a mixing study is unanswered. It seems like a clever researcher could perform some outcome studies that could give us the answer.
It is impossible for local institutions to develop reference intervals for infants or small children, given standard consent restrictions. The most authoritative set of neonate and infant ranges is provided in Andrew M, et al. Development of the hemostatic system in the neonate and young infant. Am J Pediatr Hematol Oncol 1990;12:95-104. I’ve reproduced the late Dr. Andrew’s ranges on http://uabcoag.net, a web site I developed for theUniversity of Alabama at Birmingham special coagulation laboratory. I’d suggest that for neonates and infants, any result prolonged past the upper limit should be followed up.
My audio module, Method Validation in Hemostasis, Part 2 addresses the development of a reference interval. In summary, to verify a previously defined reference interval, an institution should test 60 normal specimens that are representative of their demographic, then compute mean +/- 2 SD or a 95% confidence interval.
The most efficient way to do this is to collect normal specimens from healthy adult subjects among your institutional population over a period of about a week. We often collect from medical students, nursing students, and medical laboratory science students, as well as staff, residents, and fellows, though one could argue these populations may not be perfectly representative of our patient population. We try to have a 50-50 distribution of men and women. Centrifuge the freshly collected specimens to produce platelet-poor plasma (less than 10,000 platelets/mcl), distribute to numerous ~0.5 mL aliquots in sealed labeled freezer containers, and freeze at -70°C. For efficiency, collect enough blood to make 20 to 30 aliquots per subject. If you are freezing at -20°C, prepare 5 to 10 aliquots, as the aliquots are likely to deteriorate more rapidly.
When the time comes to prepare a new reference interval, thaw, mix and assay 1 aliquot each from ten normal subjects per day for six days, involving all your testing personnel. Collect the data, and compute the 95% confidence interval. For PTTs especially, a new reference interval is necessary with every lot change.
By the way, the American Association of Clinical Chemists, AACC, is currently engaged in a project to develop new reference intervals for many analytes in infants and neonates. Geo.