We study constraints on type-II two-Higgs-doublet models at large tan beta from CERN LEP and SLD Z-pole data and from lepton universality violation in W decay. We perform a global fit and find that, in the context of Z decay, the LEP-SLD experimental values for lepton universality violation, R-b, and A(b) all somewhat disfavor the model. Contributions from the neutral-Higgs sector can be used to constrain the scalar-pseudoscalar Higgs mass splittings. Contributions from the charged-Higgs sector allow us to constrain the charged-Higgs boson mass. For tan beta=100 we obtain the 1 sigma classical (Bayesian) bounds of m(H+/-) greater than or equal to 670 GeV (370 GeV) and 1 greater than or equal to m(h)0/m(A)0/(alpha = beta) =m(H)0/m(A)0/(alpha=0) greater than or equal to 0.68 (0.64). The 2 sigma bounds are weak. Currently, the Fermilab Tevatron experimental limits on lepton universality violation in W decay provide no significant constraint on the Higgs sector.