<p>This study examined the associations between social network types and the worsening of depressive feelings during the first and second COVID‑19 waves in Europe, with attention to gender differences and changes over time. The analysis used longitudinal data from the Survey of Health, Ageing and Retirement in Europe (SHARE), covering individuals aged 50 and older in 26 European countries and Israel. Linear probability models were estimated to assess how seven pre-pandemic confidant network types – identified through k-means clustering – were linked to self‑reported worsening of depressive feelings in each pandemic wave. Results showed that confidant networks centred on a spouse were associated with a lower risk of worsening depressive feelings. In contrast, networks composed mainly of friends were associated with a higher risk, and extended‑family networks showed no clear protective effect. Having no named confidant was not clearly associated with a greater worsening of depressive feelings. Women faced a higher risk of worsening depressive feelings than men across all confidant network types; this gender gap was larger during the first COVID‑19 wave than the second. These patterns suggest that spouse‑inclusive confidant networks offered relatively greater emotional and practical support during the acute phase of the crisis, while non‑family networks were more severely affected by distancing policies. The study highlights the protective role of close family ties within one’s core confidant network and underscores the importance of gender‑sensitive mental health strategies during global crises.</p>