Binding and maintaining cells inside microfluidic channels is a challenging task due to the potential release of cells from the channels with the flow and accompanying shear stress. In this work we optimized the binding of human B-lymphocyte cells (HR1K) inside a microfluidic channel and determined the strength of this binding under shear stress of flowing liquid. In order to determine the parameters required for a live/dead test in microfluidic devices, populations of both living and dead cells were tested separately. Channels were prepared in glass-polydimethylsiloxane hybrid chips, with a self-assembled monolayer of 3-(glycidyloxypropyl)trimethoxysilane (GPTMS) before covalently immobilizing anti-CD20 antibody. Without GPTMS linker, ~90% of the CD20-expressing cells detached at 200 μL/min (the highest flow rate studied). With GPTMS linker, the bonding method proved critical for sustained immobilization of HR1K cells under flow. Masking the channel area during plasma bonding preserves the antibody functionality; the masked surface gives 15% cell detachment at 200 µL/min compared with 80% for an unmasked surface. Sealing the chip via clamping (without plasma treatment) was similar to masked plasma treatment (20% detachment) and allowing a post-adhesion stasis time (30 min) did not significantly change the relative cell detachment for the flow rates studied. Membrane integrity and calcium spiking behaviour were measured fluorescently, and demonstrated that the live cells retained comparable functionality to unanchored cells for the duration of the flow experiments. Non-viable HR1K cells were found to detach more readily, exhibiting only 20% cell retention at 200 μL/min compared with > 80% for live cells.