Well, sorry, but your new circuit ruined the essence of that little 1 opamp design, since you lost the body ground, and killed the high-impedance input. I have been down this path as I design eeg systems which are much like ecg circuits.
In the original circuit, the battery is divided in half to create a ground voltage. A floating battery in combination with that divider provides a pretty clean pos/neg supply for the opamp. That ground also gets connected to the leg (right leg usually) and is the body ground for the circuit – without that connection to the body you will get garbage signals. You took that lead off your new circuit.
The big problem is that your two chest leads now feed into low impedances of your filters, and the millivolt signals will get lost (well, chest-1 is the desired signal and chest-2 is for noise cancellation). The original design had the main chest lead feeding into the opamp pos input which is biased to the half-supply ground via a 10 meg which is a nice high impedance (the choice of a TL072 is also probably good, as it is a fet input opamp with a much higher input impedance than an ancient bipolar like the old 741).
The circuit is amplifying the Chest-1 lead by 1+R5/R3 which is a gain of about 57. You can tweak these to change gain if needed. Now R3 needs to return to the same dc reference as the positive input which is that same half-supply ground, but they are connecting it to chest-2. This is interesting for a couple of reasons: the chest will be at about the same dc level so the opamp should get biased properly, AND it will be picking up about the same amount of 60-cycle (or 50) powerline noise as the chest-1 lead, but the opamp is subtracting that out to a large extent. This is called a common-mode signal (in-phase noise on both inputs).
I like a good sallen-and-key filter as much as the next guy, but you need to re-visit the original circuit which probably works ok. And please don’t draw a ground symbol facing any direction but down
If the opamp output just sits at the pos or neg rail, there is a dc offset that needs to be cancelled out. If it is oscillating at MHz freqs like you often see on a breadboard, add a small cap across R5. In fact, you should always have a small cap across a feedback resistor to keep the opamp from oscillating or at least to limit bandwidth – it provides a rolloff of high freqs starting at a -3dB point of 1/(2piRC); eg: a 100K with 100pF across it is a low-pass opamp circuit starting about 16KHz.
Try it in spice. Spice is great for opamp circuits – start with an ideal opamp to get the circuit topology figured out and then put real opamp models in to see how they affect the circuit. I have not yet tried the spice in kicad but have done a lot of ltspice, and many other versions going back to when it was loaded from a 9-track tape. I have learned a lot about circuits by simulating before prototyping even begins. Some spice opamp models even do a decent job with 1/f noise corners and other subtleties so you can simulate system noise floors as well as signals.
Caution: connecting things near your heart can be dangerous, and this basic circuit has no current limiting or protective considerations at all. To be completely safe you should not connect this to a human. If you choose to do so, ONLY use a battery to power this, and a battery to power the arduino/micro, and use a laptop on battery as well. No usb cable to a desktop – no connections to the ac mains anywhere…