Animal studies have shown that nickel compounds may induce pulmonary fibrosis, but so far only limited documentation in humans has been available. Radiographs of 1046 workers in a nickel refinery in Norway were read blindly and independently by three NIOSH certified B-readers, according to the ILO standards. Pulmonary fibrosis (PF) was defined as a median reading of ILO score ≥1/0 and following this criterion, 47 cases (4.5%) were identified. In logistic regression models, controlling for age and smoking, there was evidence of increased risk of PF with cumulative exposure to soluble nickel or sulfidic nickel (p = 0.04 for both). For metallic nickel a p-value of 0.07 was found. For soluble nickel there was a dose–response trend for 4 categories of cumulated exposure. In the group with the highest cumulative exposure to soluble nickel (low exposure as reference), the crude odds ratio for PF was 4.34 (95% CI 1.75–10.77). The risk adjusted for age, smoking, asbestos and sulfidic nickel was 2.24 (0.82–6.16), with a dose–response trend. The corresponding figures for sulfidic nickel were 5.06 (1.70–15.09, crude) and 2.04 (0.54–7.70, adjusted for age, smoking, asbestos and soluble nickel). However, the dose–response trend was less clear for sulfidic nickel. Controlling for estimated asbestos exposure at the refinery tended to increase the odds ratios of soluble and sulfidic nickel. This study indicates that in addition to age and smoking exposure to soluble and sulfidic nickel compounds are risk factors of PF in humans. Since the number of cases identified in this study is small and undetected confounders may have been present, further studies in other cohorts are appropriate.