OPINION OF THE COURT
When calculating the income benefit for the claimant’s work-related injury, the Administrative Law Judge (ALJ) apportioned 68% of the 21% permanent impairment rating that existed at maximum medical improvement (MMI) to a pre-existing active condition, which was non-compensable. The Court of Appeals affirmed the Workers’ Compensation Board’s decision to vacate the calculation on the ground that the ALJ should have subtracted the pre-exist-ing active impairment rating that existed immediately before the injury' from the impairment rating that existed at MMI and based the income benefit on the remainder. Appealing, the employer maintains that nothing prevented the ALJ from apportioning the impairment rating at MMI based on permissible inferences drawn from the medical evidence.
We affirm. The Board and the Court of Appeals applied the correct methodology for determining the impairment rating upon which to base income benefits. Medical opinions apportioning 63% of causation in 2007 or five-eighths of the impairment rating assigned in 2008 to a pre-existing active condition would not support a reasonable inference that the same portion of the 21% impairment rating that existed after a work-related spinal fusion in 2009 was non-work-related.
The claimant was born in 1970 and worked for the defendant-employer as a car salesman. His back problems began after a 1989 football injury that caused him to quit his high school’s team. Medical records documented his treatment for back complaints from 2001 through June 2005. The work-related injury at issue presently occurred on November 7, 2005, when the vehicle in which he accompanied a customer on a test drive was rear-ended while traveling at approximately 50 miles per hour on an interstate highway.
Dr. Gullo began treating the claimant in May 2001, at which time he reported a history of a herniated/ruptured disc since 1989. An MRI performed in July 2001 confirmed the presence of degenerative disc disease at L4-5 and L5-S1; a “tiny central protrusion” at L4-L5; a “very small paracentral disc protrusion” at L5-S1 on the left; and some crowding of the SI nerve root on the left. Dr. Gullo referred the claimant for physical therapy. He returned in May 2002 and September 2002 for complaints of back pain.
The claimant returned to Dr. Gullo again in September 2004 with recurrent back pain and pain down his right leg. An MRI performed in October 2004 revealed a new 4 mm. disc protrusion at L4-5 with *600compression of the theca and right anteri- or nerve root. The radiologist described the L5-S1 protrusion as being stable.
Dr. Yamamoto examined the claimant in October 2004 at Dr. Gullo’s request; compared the 2001 and 2004 MRIs; and described the two studies as being “basically identical.” He concluded that the claimant’s back pain resulted from lumbar disc degeneration at L4-5 and L5-S1 and “strongly” recommended both conservative treatment and daily exercise. Dr. Yama-moto found no evidence of radiculopathy and described the claimant’s condition as having been stable for over three years.
The claimant returned to Dr. Gullo in February 2005, complaining of constant back pain that flared up every two to three weeks. He also complained of “a lot more” back pain radiating into his right buttock and leg and causing weakness in the leg. Dr. Gullo continued to prescribe anti-inflammatory and pain medication but noted that the claimant might be a candidate for epidural cortisone injections.
Dr. Lockstadt began to treat the claimant in February 2005 for predominately right-sided lower back and hip pain, with occasional pain into the buttock and thigh. He interpreted the October 2004 MRI as showing a moderately-sized central disc herniation at L4-5, with no apparent nerve root compression; a smaller left-sided bulge at L5-S1; and some mild to moderate facet joint arthritis at L4-5 and L5-S1. Dr. Lockstadt interpreted lumbar spine x-rays performed in February 2005 as showing a pattern of instability at L4-5 with Grade I retrolisthesis. Physical examination revealed no muscle spasm or scoliosis, but the claimant did experience pain on flexion, extension, and straight leg raising. Dr. Lockstadt attributed the majority of the claimant’s pain to the L4-5 level and recommended a series of epidural steroid injections followed by physical therapy. He also considered the claimant to be a candidate for a transforaminal discectomy at L4-5 and L5-S1.
Dr. Lockstadt noted in April 2005 that injections performed recently at the L4-5 and L3-4 levels had improved the claimant’s leg symptoms dramatically although he continued to experience some mechanical low back pain. The claimant began physical therapy sometime thereafter. He had a third lumbar injection in June 2005, at which time he reported that his right leg symptoms had essentially disappeared but that he continued to have some moderate low back pain.
The claimant returned to Dr. Lockstadt on November 7, 2005, complaining that his back pain had worsened since the motor vehicle accident two days earlier. Dr. Lockstadt examined him and noted as follows:
Based on his pain, which was still quite severe prior to the motor vehicle accident, I think it is worthwhile going ahead and updating the MRI scan. We will see if he would be a candidate for a disc replacement. The epidurals did not help him with his back pain previously. Therefore I do not think they would help this time either as he has a painful disc, which has been markedly worsened by the motor vehicle accident.
An MRI performed on November 17, 2005 revealed a posterior disc bulge with a more focal posterocentral disc extrusion measuring “approximately 9 mm. in diameter” and facet arthropathy at L4-5, producing stenosis in the spinal canal, right lateral recess, and bilateral neural fora-men. The scan also revealed a posterior disc bulge with a more focal left posterior parasagittal disc protrusion and facet ar-thropathy at L5-S1, producing stenosis in the left lateral recess and bilateral neural foramen.
Dr. Lockstadt concluded on November 21, 2005 that the claimant was not a candidate for the previously-considered trans-*601foraminal discectomy at L4-5 and L5-S1 based on the extent of the disc degeneration shown on the most recent MRI. Moreover, a sequester fragment at L4-5 had moved inferiorly since the previous scan. He concluded that the only procedure likely to help the claimant at present would be a two-level fusion.
Dr. Lockstadt opined in January 2006 that the motor vehicle accident appeared to have caused an enlargement of the L4-5 herniation with further lateralization towards the right, basing his opinion on a comparison of the 2004 and 2005 MRI results. He stated that the accident “likely caused an aggravation of the L4-5 herniation and a subsequent increased fluid buildup in the L4-5 inter spinous space” and that the scans revealed only a minimal increase in the amount of degeneration at L5-S1.
A follow-up report by Dr. Lockstadt in March 2006 stated:
The change that I have noted is at the L4-5 level. In 2004 he had a central herniation, which occupied approximately 50% of the transverse disc space. Compared to the scan which was done just a month ago, it shows that the herniation has increased, involving approximately 75% of the transverse distance of the disc space and an increase in posterior herniation. There is also now a small amount of lateralization towards the light. This therefore does correlate with his symptoms.
Dr. Guarnaschelli evaluated the claimant in May 2006 and concluded that the work-related accident was a substantial factor in aggravating the claimant’s pre-existing back condition. He did not assign a permanent impairment rating but would attribute half of the claimant’s impairment to his pre-existing active condition. Dr. Guarnaschelli testified when deposed in December 2007 that the claimant was approaching MMI rapidly as of May 2006 and was not a surgical candidate at that time but that he might require lumbar surgery in the future. He opined based on the pre- and post-accident MRI scans that the accident probably would cause a harmful change that was permanent.
Dr. Sexton reviewed the claimant’s medical records in August 2006. He apportioned 90% of the need for low back treatment to the pre-existing condition and 10% to the work-related accident. He opined, moreover, that the accident caused no increase in the claimant’s “permanent partial impairment.”
Dr. Kriss evaluated the claimant in December 2007. He noted the claimant’s history of “very chronic, very active, and very significant low back pain prior to November 5, 2005” but also noted that the accident on that date “does not appear to have been a typical or minor ‘rear-end’ collision.” He opined that the pre- and post-accident MRI scans clearly showed a permanent harmful change as a direct result of the accident. Turning to the question of how to apportion the claimant’s impairment between the work-related and non-work-related causes, he noted that he viewed the 20-year history of significant back pain in a “very young man” to be “more impressive” than changes found on the most recent MRI scan. He then apportioned “63% of the total lumbar causation” to pre-existing degenerative disc disease and “37% of the total lumbar causation” to the work-related accident.
Dr. Kriss reported in December 2008, after reviewing additional medical records, that the claimant’s present lumbar condition warranted an 8% permanent impairment rating. His condition before the accident warranted a 5% impairment rating. Subtracting the pre-accident impairment rating from the present rating, Dr. Kriss attributed a 3% impairment rating to the effects of the work-related accident.
*602Dr. Lockstadt continued to provide conservative treatment for the claimant’s symptoms through March 2009, at which time he performed lumbar fusion surgery. Dr. Lockstadt noted in December 2009 that the claimant had a successful L4-5 and L5-S1 fusion “for back pain that had a mild pre-existing component made worse by the motor vehicle accident;” noted that his back pain had shown “dramatic improvement;” and released him to return to work. He testified when deposed in March 2010 that the claimant’s pre-acci-dent permanent impairment rating would have been no more than 5% and that his present impairment rating was 21%. He opined that the post-accident worsening of the claimant’s condition and the need for the lumbar fusion probably resulted from the accident.
Relying on Drs. Lockstadt and Guarnas-chelli, the ALJ found that the work-related accident caused a permanent harmful change in the claimant’s low back as well as the need for surgery. The ALJ found that the claimant’s present condition warranted a 21% permanent impairment rating. Relying on the apportionment of causation made by Dr. Kriss in December 2007, the ALJ attributed 63% of the impairment rating to the pre-existing active degenerative condition, which was not compensable. The ALJ found the remaining 37% of the 21% impairment rating, i.e., 7.77%, to be compensable for the purpose of awarding income benefits.
The Board determined that the ALJ erred as a matter of law by relying on Dr. Kriss’s apportionment of causation to base income benefits on a permanent impairment rating that no medical expert assigned using the AMA Guides to the Eval-nation of Permanent Impairment (Guides). Noting that the only impairment rating assigned at MMI following the surgery was 21% and that both Dr. Lock-stadt and Dr. Kriss assigned a 5% pre-accident impairment rating, the Board determined that the evidence compelled the claimant’s award to be based on a 16% impairment rating.
Affirming the decision, the Court of Appeals rejected the argument that we face presently. The employer maintains that the ALJ did not err but exercised a fact-finder’s discretion to infer reasonably that a progression of the pre-existing degenerative condition contributed in the same proportion to causing the claimant’s present impairment rating as it did to causing the impairment rating that Dr. Kriss assigned in 2008. We disagree.
Since 1996 Chapter 342 has required permanent partial disability awards to be calculated based on the product of a permanent impairment rating that is assigned using the AMA Guides and a corresponding statutory factor.1 The impairment rating that a pre-existing, active condition warrants must be excluded from the impairment rating upon which benefits are based because only work-related impairment is compensable. In Kentucky River Enterprises, Inc. v. Elkins2 the court rejected the notion that an ALJ may base benefits on an impairment rating that is not assigned by a medical expert using the criteria set forth in the Guides.
We do not agree with the employer’s argument that Dr. Kriss’s opinions apportioning 63% of the condition present in 2007 and five-eighths of the impairment rating assigned in 2008 to a non-work-related cause3 supported a reasonable in*603ference that 68% of the impairment rating present after the claimant’s work-related spinal fusion in 2009 was also non-work-related. Dr. Kriss last examined the claimant in 2007 and last reviewed any of his medical records in December 2008. Neither he nor any other medical expert testified that a progression of the preexisting degenerative condition contributed in the same proportion to the impairment rating present after the 2009 fusion surgery, which the ALJ found to be work-related. Moreover, no medical expert testified that the Guides authorize the apportionment of an impairment rating in the manner employed by Dr. Kriss.
The Board and the Court of Appeals determined correctly that the ALJ erred by failing to subtract the 5% impairment rating that physicians assigned based on the back condition as it existed immediately before the claimant’s accident from the 21% impairment rating that existed when he reached MMI after the work-related lumbar fusion. The ALJ must award income benefits based on the 16% impairment rating that remains.
The decision of the Court of Appeals is affirmed.
All sitting. All concur.. See KRS 342.0011(1 l)(b), (35), and (36); KRS 342.730(l)(b).
. 107 S.W.3d 206 (Ky.2003).
.The employer reasons that Dr. Kriss assigned a 5% pre-existing impairment rating and an 8% impairment rating in 2008. Thus, five-eighths of the post-accident impairment *603rating was non-work-related. The employer notes that five-eighths equals 63%.