Table 2.

Regression analyses on dataset 1 with additional interaction terms coefficients

Group	Stat. Sign. Coeff.	Est.	Std.Err	t-value	Pr(>|t|)
PD_MHPG	Intercept	12.333	2.892	4.264	0.000594***
	n	2.611	3.587	0.728	0.477173
	method LC-ED	3.737	11.341	0.329	0.746052
	method MF	−44.624	51.843	−0.861	0.402089
	method GC	−3.461	26.520	−0.130	0.897799
	method GLC	−23.880	16.270	−1.468	0.161562
	csfvol	3.736	3.522	1.060	0.304664
	age	5.870	4.270	1.375	0.188225
	severity	5.933	9.209	0.644	0.528502
	ypd	4.228	8.358	0.506	0.619799
	age* severity	9.412	16.634	0.566	0.579335
	age* ypd	5.020	7.299	0.688	0.501489
PD_NA	Intercept	128.713	54.928	2.343	0.0472*
	n	−8.559	39.928	−0.214	0.8356
	method LC-ED	205.982	127.914	1.610	0.1460
	method RM	185.784	114.636	1.621	0.1438
	csfvol	30.076	59.731	0.504	0.6282
	age	90.557	87.155	1.039	0.3292
	severity	38.585	47.864	0.806	0.4435
	ypd	−7.809	40.281	−0.194	0.8511
	age*severity	−198.496	133.884	−1.483	0.1765
	age*ypd	−2.316	104.947	−0.022	0.9829

A further regression was performed on both NA and MHPG levels to explore the interaction between the variable ‘age’ and the variables ‘severity’ and ‘years post diagnosis’, previously excluded due to their absence for more than half of the participants. For both variables, NA and MHPG, the tested regression model was then n + ‘method’ + ‘csfvol’ + ‘age’*‘severity’ + ‘age’*‘ypd’. Coeff = coefficient; ED = electrochemical detection; Est = parameter estimates; GC = gas chromatography; GLC = gas liquid chromatography; LC = liquid chromatography; MF = mass fragmentography; MHPG = 3-methoxy-4-ydroxyphenylglycol; n = sample size; NA = noradrenaline; PD = Parkinson’s Disease; Pr(>|t|) = P-value associated with the t statistic; RM = radioenzymatic methods; Stat.Sign.Coeff = statistically significant coefficient; Std.Err = standard error. Significance levels are indicated by asterisks (*P < 0.05, ***P < 0.001).