I would do something like this:
% Sampling
fs = 1000;
Ts = 1/fs;
% Time vector
t = 1:Ts:10-Ts;
% Signal
f = 1; % Frequency [Hz]
a = 1; % Amplitude
signal = sin(2*pi.*t.*f); % Sample sinusoidal signal. Your ECG signal goes here.
% Noise
fNoise = 50; % Frequency [Hz]
aNoise = 0.25; % Amplitude
noise = aNoise*sin(2*pi.*t.*fNoise);
% Signal + Noise
signalNoise = signal + noise;
% Plots
figure();
subplot(3,1,1);
plot(t, signal);
xlabel('Time [s]');
ylabel('Amplitude');
title('Original signal');
subplot(3,1,2);
plot(t, noise);
xlabel('Time [s]');
ylabel('Amplitude');
title('Noise');
subplot(3,1,3);
plot(t, signalNoise);
xlabel('Time [s]');
ylabel('Amplitude');
title('Original signal + Noise');
The trick is that your noise signal must have the same length than your ECG signal. If you do it this way, you can warranty that. Try it and let me know if it works ;-)
