Isoniazid was subjected to different ICH prescribed stress conditions of thermal stress, hydrolysis, oxidation and photolysis. The drug was stable to dry heat (50 and 60 degrees C). It showed extensive decomposition under hydrolytic conditions, while it was only moderately sensitive to oxidation stress. The solid drug turned intense yellow on exposure to light under accelerated conditions of temperature (40 degrees C) and humidity (75% RH). In total, three major degradation products were detected by LC. For establishment of stability-indicating assay, the reaction solutions in which different degradation products were formed were mixed, and the separation was optimized by varying the LC conditions. An acceptable separation was achieved using a C-18 column and a mobile phase comprising of water:acetonitrile (96:4, v/v), with flow rate and detection wavelength being 0.5 ml min(-1) and 254 nm, respectively. The degradation products appeared at relative retention times (RR(T)) of 0.71, 1.34 and 4.22. The validation studies established a linear response of the drug at concentrations between 50 and 1000 microg ml(-1). The mean values (+/-R.S.D.) of slope, intercept and correlation coefficient were 35,199 (+/-0.88), 114,310 (+/-4.70) and 0.9998 (+/-0.01), respectively. The mean R.S.D. values for intra- and inter-day precision were 0.24 and 0.90, respectively. The recovery of the drug ranged between 99.42 and 100.58%, when it was spiked to a mixture of solutions in which sufficient degradation was observed. The specificity was established through peak purity testing using a photodiode array detector. The method worked well on application to marketed formulation of isoniazid, and a fixed-dose combination containing isoniazid and ethambutol HCl. It was even extendable to LC-MS studies, which were carried out to identify the three degradation products. The m/z values of the peaks at RR(T) 0.71 and RR(T) 1.34 matched with isonicotinic acid and isonicotinamide, respectively. The product appearing at RR(T) 4.22 was isolated using preparative LC-MS, and turned out to be a yellow compound that was identified as isonicotinic acid N'-(pyridyl-4-carbonyl)-hydrazide based on mass, FTIR and (1)H/(13)C NMR spectral data. The same was indicated to be responsible for discolouration of isoniazid bulk drug substance and formulations, which is a familiar problem. The mechanism of formation of the said compound is outlined.