Thrombotic complications are common in Coronavirus disease 2019 (COVID-19) patients, with pulmonary embolism (PE) being the most frequent. Randomised trials have provided inconclusive results on the optimal dosage of thromboprophylaxis in critically ill COVID-19 patients. We utilized data from the multicentre CAPACITY-COVID patient registry to assess the effect of differential application of Low Molecular Weight Heparin (LMWH) dose protocols on PE and in-hospital mortality risk in critically ill COVID-19 patients. An instrumental variable analysis was performed to estimate the intention-to-treat effect, utilizing differences in thromboprophylaxis prescribing behaviour between hospitals. We included 939 patients with PCR confirmed SARS-CoV-2 infection from 34 hospitals. Two-hundred-and-one patients (21%) developed a PE. The adjusted cause-specific HR of PE was 0.92 (95% CI: 0.73-1.16) per doubling of LMWH dose. The adjusted cause-specific HR for in-hospital mortality was 0.82 (95% CI: 0.65-1.02) per doubling of LMWH dose. This dose-response relationship was shown to be non-linear. To conclude, this study did not find evidence for an effect of LMWH dose on the risk of PE, but suggested a non-linear decreased risk of in-hospital mortality for higher doses of LMWH. However, uncertainty remains, and the dose-response relationship between LMWH dose and in-hospital mortality needs further investigation in well-designed studies.