The incidence of regression after the non-surgical treatment of symptomatic lumbar disc herniation: a systematic review and meta-analysis

Table 3 Subgroup analyses of the regression measurement, time period, region, study type, LDH level, imaging method and patient count

	Included studies (n)	Number of patients		Incidence (95% CI)	P value^*
	Included studies (n)	With Regression	Total	Incidence (95% CI)	P value^*
Measurement	38	1425	2219	63%(0·49–0·77)	0.01
Qualitative	25	627	1314	54% (0·37–0·70)
Quantitative	13	798	905	81% (0·69–0·91)
Time period	38	1425	2219	63%(0·49–0·77)	0.87
Before 2000	15	353	565	65% (0·55–0·75)
2000–2009	11	280	530	57% (0·29–0·83)
2010–2019	12	792	1124	66% (0·35–0·91)
Region	38	1425	2219	63%(0·49–0·77)	0.97
Asia	19	879	1309	63% (0·40–0·83)
Europe	14	459	756	65% (0·43–0·85)
North America	5	87	154	60% (0·22–0·92)
Study type	30^a	1315	1965	65%(0·48–0·80)	0.14
RCT	5	125	421	37% (0·00–0·88)
Prospective	20	588	882	67% (0·57–0·77)
Retrospective	5	602	662	84% (0·65–0·97)
LDH level	29^b	1241	1667	72% (0·58–0·84)	0.17
Single	22	1103	1340	78% (0·67–0·87)
Single/multiple	7	138	327	51% (0·17–0·86)
Imaging method	37^c	1416	2208	63%(0·48–0·76)	0.97
CT	8	202	343	63% (0·46–0·79)
MRI	29	1214	1865	63% (0·45–0·79)
Counting^d	38	1425	2219	63%(0·49–0·77)	0.33
A	30	1279	2018	61% (0·44–0·77)
B	8	146	201	72% (0·59–0·83)

*P value is from the test for subgroup differences (random effects model)
^aEight studies did not report study type
^bNine studies did not report LDH level
^cOne study used CT at basline and MRI at follow-up
^dCounting. A. The number of patients with regression was reported and was extracted from the publication. B. For studies that recorded the number of patients by the regression proportion or size interval but did not define the interval of non-regression or report the number of patients without regression, we regarded the lowest interval as the no regression range, and the number of patients outside of this interval was considered the number of patients with regression

ISSN: 1471-2474