This paper explores a star-of-star topology for an internet-of-things (IoT) network using mega low Earth orbit constellations where the IoT users broadcast their sensed information to multiple satellites simultaneously over a shared channel. The satellites use amplify-and-forward relaying to forward the received signal to the ground station (GS), which then combines them coherently using maximal ratio combining. A comprehensive outage probability (OP) analysis is performed for the presented topology. Stochastic geometry is used to model the random locations of satellites, thus making the analysis general and independent of any constellation. The satellites are assumed to be visible if their elevation angle is greater than a threshold, called a mask angle. Statistical characteristics of the range and the number of visible satellites are derived for a given mask angle. Successive interference cancellation (SIC) and capture model (CM)-based decoding schemes are analyzed at the GS to mitigate interference effects. The average OP for the CM-based scheme, and the OP of the best user for the SIC scheme are derived analytically. Simulation results are presented that corroborate the derived analytical expressions. Moreover, insights on the effect of various system parameters like mask angle, altitude, number of satellites and decoding order are also presented. The results demonstrate that the explored topology can achieve the desired OP by leveraging the benefits of multiple satellites. Thus, this topology is an attractive choice for satellite-based IoT networks as it can facilitate burst transmissions without coordination among the IoT users.