Tyrosine kinase inhibitors (TKIs) that target epidermal growth factor receptor (EGFR) have dramatically improved progression-free survival in non-small-cell lung cancer (NSCLC) patients who carry sensitizing EGFR-activating mutations and in patients with breast and pancreatic cancers. However, EGFR-TKIs are associated with significant and disabling undesirable effects that adversely impact on quality of life and compliance. These effects include dermatological reactions, diarrhoea, hepatotoxicity, stomatitis, interstitial lung disease and ocular toxicity. Each individual EGFR-TKI is also associated with additional adverse effect(s) that are not shared widely by the other members of its class. Often, these effects call for dose reduction, treatment discontinuation or pharmacotherapeutic intervention. Since dermatological effects result from on-target effects on wild-type EGFR, rash is often considered to be a biomarker of efficacy. A number of studies have reported better outcomes in patients with skin reactions compared with those without. This has led to a 'dosing-to-rash' strategy to optimize therapeutic outcomes. Although conceptually attractive, there is currently insufficient evidence-based support for this strategy. While skin reactions following EGFR-TKIs are believed to result from an effect on wild-type EGFR, their efficacy is related to effects on mutant variants of EGFR. It is noteworthy that newer EGFR-TKIs that spare wild-type EGFR are associated with fewer dermatological reactions. Furthermore, secondary mutations such as T790M in exon 20 often lead to development of resistance to the clinical activity and efficacy of first- and second-generation EGFR-TKIs. This has stimulated the search for later-generations of EGFR-TKIs with the ability to overcome this resistance and with greater target selectivity to spare wild-type EGFR in expectations of an improved safety profile. However, available data reviewed herein indicate that not only are these newer agents associated with the aforementioned adverse effects typical of earlier agents, but they are also susceptible to resistance due to tertiary mutations, most frequently C797S. At least three later-generation EGFR-TKIs, canertinib, naquotinib and rociletinib, have been discontinued from further development in NSCLC following concerns about their safety and risk/benefit.