Clinical Impact of New Reference Intervals for the Roche Prolactin II Immunoassay

Abstract

Context: The Roche prolactin immunoassay is used throughout the world. It reports higher values than the Siemens immunoassay but the manufacturer-defined reference intervals are similar. Patient results are often above the Roche upper limit but within the Siemens interval, causing diagnostic confusion.

Objective: Establish new reference intervals for the Roche and Siemens prolactin immunoassays.

Methods: We established new reference intervals for the Roche and Siemens immunoassays using 374 specimens from healthy outpatients. We performed chart review for unnecessary testing and treatment for 298 patients in a 6-month period with at least 1 Roche prolactin value above the manufacturer-defined upper limit and below our new upper limit.

Results: The new upper limit for the Roche assay was 37.8 ng/mL (females) and 22.8 ng/mL (males). The manufacturer-defined limits were 23.3 ng/mL and 15.2 ng/mL, respectively. New intervals for the Siemens assay matched the manufacturer. No cases of clinically significant pathophysiologic prolactin excess were identified in patients with values between the manufacturer-defined upper reference limit and our new Roche upper limit. Unnecessary further evaluation in these patients included 459 repeat prolactin measurements, 57 macroprolactin measurements, 39 magnetic resonance imaging studies, and 28 endocrine referrals. Eleven patients received dopamine agonists. The minimum cost of excess care using Medicare reimbursement rates was $34 134, with substantially higher amounts billed to patients and their insurance providers.

Conclusion: Adoption of new upper reference limits for the Roche prolactin assay of 37.8 ng/mL (females) and 22.8 ng/mL (males) would not delay diagnosis or necessary intervention in patients with clinically significant pituitary tumors but would reduce unnecessary evaluation in patients without pathophysiologic prolactin excess.

Keywords: Prolactin; bias; excess care; immunoassay; reference interval.

Figure 1.

Comparison of concurrent prolactin values generated using the Siemens and Roche immunoassays on the same set of remnant clinical plasma or serum specimens. Female, female patients not prescribed ethinyl estradiol-containing contraceptives; Female EE, female patients prescribed ethinyl estradiol-containing contraceptives; Male, male patients. Linear regression equations: female: y = 1.52 × –0.098, R² 0.923; female EE: y = 1.57 × –0.558, R² = 0.984; male: y = 1.63 × –1.139, R² = 0.976.

Figure 2.

Roche prolactin values classified by daily ethinyl estradiol dose in the group prescribed ethinyl estradiol-containing contraceptives. t-test P values 15 vs 20 µg/day: 0.212, 15 vs 30 µg/day: 0.149, 15 vs 35 µg/day: 0.321, 20 vs 30 µg/day: 0.794, 20 vs 35 µg/day: 0.717, 30 vs 35 µg/day: 0.525.