Secondary claim guide

Nexus Letter for Back Pain Secondary to a Service-Connected Knee

Medically reviewed by the Patriot Path Medical Team

Licensed MD reviewers • Last updated: November 7, 2026

Patriot Path Medical Team → • Our review process →

✓ Medically Reviewed

VA can service connect a back condition secondary to a service-connected knee under 38 CFR 3.310 when the evidence links them, by causation or by aggravation. A limp is how the knee is said to reach the back, not a measure of the back rating, and a grant can be rated 0 percent or leave your combined rating unchanged.

Three points belong up front. You may not need to buy a letter: once service-connected arthritis has been medically determined to be degenerative, not post-traumatic, VA's manual says each joint the disease reaches, verified by x-ray, is subject to service connection without a medical opinion, absent an intervening cause, and it applies that principle to the spine (M21-1 V.iii.1.C.2.b). That rule is narrower than it sounds, it is not secondary service connection, and we could not measure whether VA applies it to the spine today. No study we found shows that knee problems cause back problems: the research is limited and runs in both directions, so the case rests on your own records. And the Board grants we read on this pairing granted service connection, not a percentage, each "binding only with regard to the specific case decided" under 38 CFR 20.1303.

An illustrative example

A veteran is service connected for a right knee injured in a fall during training, rated 10 percent for limited motion, his only rating. His treatment notes have recorded a limp for years, an MRI shows degenerative disc disease in his lower back, and his forward flexion measures 50 degrees. A VA examiner wrote that the knee and the back are separate parts of the body. Because the knee is service connected for an injury, not for degenerative arthritis, the manual's arthritis rule does not reach his back, and the link needs medical evidence.

Suppose a physician's opinion addresses his gait history and VA grants the back as secondary to the knee. On causation, 50 degrees of flexion fits the spine formula's 20 percent row, and Table I combines 20 with 10 to 28, which rounds to 30. If instead a 2015 record had put his back at the 10 percent level before the limp worsened it, an aggravation grant would pay the difference, 10 percent, and 10 with 10 is 19, which rounds to 20. Had both records fallen at the 20 percent level, with the evidence showing some worsening, VA would grant service connection at 0 percent. The back's numbers come from its own measurements, not from the limp.

This veteran is invented to show how the rules turn assumed facts into numbers. He is not a client or a case result, and nothing here predicts what VA will decide on a real claim.

Book A Consultation See nexus letter pricing

Do You Need a Nexus Letter if Your Knee Arthritis Is Service Connected?

Possibly not, and VA's own manual is the reason: M21-1 has a rule for degenerative arthritis that reaches one joint after another, and its Note carries the rule to the spine. It is narrower than it sounds.

When 38 CFR 4.71a, DC 5003 is assigned for an arthritic condition, every joint or group of joints affected by this disease is subject to service connection (SC) in the absence of an intervening cause.M21-1 V.iii.1.C.2.b (topic change date February 8, 2021)

If it has been medically determined that the Veteran has service-connected (SC) degenerative arthritis, and not post-traumatic arthritis, there is no need for a medical opinion to establish SC for each affected joint.M21-1 V.iii.1.C.2.b (topic change date February 8, 2021)

Once degenerative arthritis has been clinically diagnosed and SC has been properly established for one affected joint, establish SC for and evaluate each joint as it becomes affected, with x-rays used to verify involvement of individual joints.M21-1 V.iii.1.C.2.b (topic change date February 8, 2021)

Note: In evaluating arthritis of the spine, the principles for establishing SC for joints affected by the subsequent development of degenerative arthritis (as contemplated under 38 CFR 4.71a, DC 5003) are not dependent on the choice of DC.M21-1 V.iii.1.C.2.b (topic change date February 8, 2021)

What the rule is not

It is not 38 CFR 3.310, and it is not secondary service connection: the block does not mention 3.310, gait or the knee. It is a rule about one disease, degenerative arthritis rated under DC 5003, reaching more joints.

Its worked example runs spine first: degenerative arthritis of the spine service connected under DC 5242, then the right elbow "with no intercurrent cause noted" (the rule's own word is "intervening"), where service connection "may be established". It is not a knee case; a knee-to-spine reading rests on the rule sentence and the Note, with 38 CFR 4.45(f) treating the lumbar vertebrae as a group of minor joints.

Where the rule stops

It applies where it "has been medically determined" that the arthritis is degenerative "and not post-traumatic". Where the type is unclear, VA is to "obtain a medical opinion to determine the etiology" (V.iii.1.C.3.b).

A knee hurt in service goes to DC 5010: "When there is evidence showing a history of in-service trauma to a joint, utilize 38 CFR 4.71a, DC 5010 for rating purposes" (V.iii.1.C.2.c), even where the evidence "diagnoses degenerative or osteoarthritis." Arthritis due to trauma "affects only the joint that received trauma" (V.iii.1.C.3.a).

Before February 7, 2021, "post-traumatic arthritis (previously called arthritis, due to trauma) was evaluated under 38 CFR 4.71a, DC 5003" (V.iii.1.C.2.d), so a DC 5003 code on an older decision does not by itself show degenerative arthritis.

The spine needs arthritis on x-ray: "X-ray evidence of bone changes must be present to substantiate a diagnosis of arthritis" (V.iii.1.C.2.a); disc changes without radiographic arthritic changes are "not indicative of arthritis" (V.iii.1.B.3.a). DC 5242 also covers "degenerative disc disease other than intervertebral disc syndrome" and points to "either DC 5003 or 5010", so a 5242 code is not by itself spinal arthritis, which can itself be post-traumatic.

Read the whole knee code: arthritis rated on motion is hyphenated, as in 5003-5261 (V.iii.1.C.4.a), or 5010 with the joint's code, so a bare 5260 or 5261 settles nothing either way.

What the Board decisions show

In the Board decisions we read, including eight decided after the topic's February 8, 2021 change date, the Board weighed or sought medical opinions on this pairing, apart from one denial because the knees were not service connected (24001968), and did not cite C.2.b or DC 5003. The nearest fact pattern, 21066807, had both knees service connected "with degenerative arthritis" and degenerative lumbar findings on x-ray, and was granted on private opinions; but its knees also carried a meniscectomy and a ligament repair, and it states no diagnostic code. Whether regional offices or the Board apply C.2.b to the spine today, we could not measure. Each decision is binding only with regard to the specific case decided (38 CFR 20.1303).

Where a letter does the work

Everywhere else, the link needs a 38 CFR 3.310 opinion. For a knee injured in service and rated under DC 5010, the opinion has to deal with the injury history, since history alone "may be the only distinguishing element for a medical diagnosis of arthritis due to trauma" (V.iii.1.C.3.a).

The same goes for a knee under a code that is not arthritis (DC 5256 to 5259, 5262, 5263, or by analogy), argued from its own mechanism, such as instability, locking or a prescribed brace; a back that is not arthritis on x-ray, such as a strain (DC 5237), a herniated disc pressing on or irritating a nerve root (DC 5243), spondylolisthesis (DC 5239) or disc disease without arthritic changes; a later injury to place in the timeline; any aggravation theory; and a negative VA opinion.

Which route your file is on

Which route your file is on
Your knee is service connected as	Your back shows	Route	What the claim needs
Degenerative arthritis under DC 5003, medically determined not post-traumatic	Arthritis of the spine on x-ray	VA's manual arthritis rule (M21-1 V.iii.1.C.2.b)	Per the manual, no medical opinion, absent an intervening cause; the rating decision's full code and the spine x-ray
Post-traumatic arthritis under DC 5010, after an in-service injury	Any acquired back diagnosis	Secondary service connection, 38 CFR 3.310	A medical opinion on causation or aggravation that deals with the injury history
A code that is not arthritis, such as DC 5257, 5258 or 5259	Any acquired back diagnosis	Secondary service connection, 38 CFR 3.310	A medical opinion built on the knee's own mechanism
Degenerative arthritis under DC 5003	A diagnosis that is not arthritis on x-ray, such as strain or disc disease	Secondary service connection, 38 CFR 3.310	A medical opinion; disc wear without arthritic changes does not meet the manual's rule
Any knee code	Back trouble documented before the knee affected it	Aggravation, 38 CFR 3.310(b)	A baseline from the earliest records and an opinion on the worsening

The rows describe routes, not outcomes.

So we promise neither the shortcut nor its absence. If your file fits the rule, you can name it in your claim. If your knee arthritis followed an injury in service, your back diagnosis is not arthritis on x-ray, or VA has already said no, the link needs medical evidence, and that is where a letter does its work.

How Does VA Decide Back Pain Secondary to a Knee?

Outside the arthritis rule, a back claim through the knee runs on 38 CFR 3.310, as VA's manual applies it. Each Board decision this section names is binding only with regard to the specific case decided (38 CFR 20.1303).

Causation: 38 CFR 3.310(a)

38 CFR 3.310(a) reads: "Except as provided in § 3.300(c), disability which is proximately due to or the result of a service-connected disease or injury shall be service connected. When service connection is thus established for a secondary condition, the secondary condition shall be considered a part of the original condition."

VA's manual words the test as but-for: causation means the back is "the result of, or would not have occurred but for, an SC disability" (M21-1 V.ii.2.D.1.b), because in Spicer v. McDonough, 61 F.4th 1360 (Fed. Cir. 2023), the Federal Circuit "held that the language “resulting from” in 38 U.S.C. 1110 requires a standard of “but-for causation” and that a stricter proximate causation standard is inconsistent with the statute" (Block 1.a note). In A25015277 (2025), the Board wrote that the but-for standard in a secondary service-connected case "is not limited to a single cause and effect, but rather contemplates multi-causal links, including action and inaction".

Aggravation: 38 CFR 3.310(b)

Paragraph (b) covers a worsening of a nonservice-connected back that is due to the knee "and not due to the natural progress of the nonservice-connected disease", on one condition: "However, VA will not concede that a nonservice-connected disease or injury was aggravated by a service-connected disease or injury unless the baseline level of severity of the nonservice-connected disease or injury is established by medical evidence created before the onset of aggravation or by the earliest medical evidence created at any time between the onset of aggravation and the receipt of medical evidence establishing the current level of severity of the nonservice-connected disease or injury."

The regulation still prints natural progress twice, but VA's manual has set it aside: "In accordance with the Spicer holding, the Department of Veterans Affairs (VA) will apply the broader but-for standard and will no longer consider natural progress of the NSC disability when deciding a claim based on aggravation of an NSC disability by an SC disability" (M21-1 V.ii.2.D.1.a, topic change date May 1, 2026). That is VA's stated practice, not a change to 3.310(b). The baseline deduction survives (Block 1.g), and "Permanent worsening of an NSC disability by an SC disability is not required to establish SC on the basis of aggravation under 38 CFR 3.310(b)" (Block 1.a).

The general rules for every secondary pairing, not only the knee and the back, are in our guide to nexus letters for secondary conditions.

The standard: 38 CFR 3.102

VA does not ask for proof. Under 38 CFR 3.102, when "a reasonable doubt arises regarding service origin, the degree of disability, or any other point, such doubt will be resolved in favor of the claimant", and "By reasonable doubt is meant one which exists because of an approximate balance of positive and negative evidence which does not satisfactorily prove or disprove the claim."

The phrase "at least as likely as not" is not in 3.102; it is how VA opinions and Board remands word that balance, glossed in A25015277 as "(an approximate balance of positive and negative evidence, i.e., nearly equal)".

The knee must be service connected, though not first

The back claim runs through a service-connected knee. In 24001968 the Board wrote: "Since the Board has denied the Veteran's claims for service connection for left and right knee osteoarthritis, the Veteran's claim for secondary service connection is also denied as a matter of law."

Timing is a different question: "To establish causation, the primary disability need not be SC, or even diagnosed, at the time the secondary disability is incurred" (M21-1 V.ii.2.D.1.b).

The back diagnosis must be a disease or injury

38 CFR 3.303(c) ends: "Congenital or developmental defects, refractive error of the eye, personality disorders and mental deficiency as such are not diseases or injuries within the meaning of applicable legislation." 38 CFR 4.9 is its Part 4 companion. VA's manual names spine defects to note, especially spondylolysis, spina bifida, unstable or exaggerated lumbosacral joints or angle, and incomplete sacralization, and cautions: "Do not automatically classify spondylolisthesis as a congenital condition, although it is commonly associated with a congenital defect" (M21-1 V.iii.1.F.1.h; "commonly" is the manual's word). In 21046103 the Board found service connection for "mild dextroscoliosis/transitional vertebra L5" precluded in that case and granted the acquired lower back disability.

Pain with no named diagnosis is harder. In 18154311, citing Saunders v. Wilkie, 886 F.3d 1356 (Fed. Cir. 2018), the Board asked the examiner: "if there is no diagnosed disability, is it at least as likely as not that any pain reaches the level of a functional impairment of earning capacity?" That is a remand question, not a grant.

A hip, ankle or foot that changed how you walk can support a back claim under the same regulation, 38 CFR 3.310, but the knee-specific rules on this page, such as the knee's arthritis codes, do not carry over directly.

This page runs one way, from a service-connected knee to the back. For a back condition that has affected a knee, or one knee overloading the other, see our knee page.

Does a Limp From Your Knee Raise Your Back Rating?

Not by itself. The limp is how the knee is said to reach the back; it is not what the back is rated on.

muscle spasm or guarding severe enough to result in an abnormal gait or abnormal spinal contour such as scoliosis, reversed lordosis, or abnormal kyphosis38 CFR 4.71a, General Rating Formula, one option in the 20 percent row

muscle spasm, guarding, or localized tenderness not resulting in abnormal gait or abnormal spinal contour38 CFR 4.71a, General Rating Formula, one option in the 10 percent row

Each row lists alternatives joined by "or", among them thoracolumbar flexion ranges and combined-motion limits, so the gait wording is one option in each row, not the requirement for either.

Both options tie the gait to spasm or guarding in the spine, "severe enough to result in" it at 20 percent and "not resulting in abnormal gait" at 10, and neither row mentions the knee, a limp or another joint. A limp from the knee is not spasm or guarding in the back, so it does not meet the 20 percent option, however long it has lasted.

The limp sits on the knee's side of the ledger: 38 CFR 4.45(f) says "Instability of station, disturbance of locomotion, interference with sitting, standing and weight-bearing are related considerations", and 38 CFR 4.40 makes functional loss part of examining the rated joint. The six VA manual articles we read have no altered-gait rule (gait, limp, antalgic and altered appear zero times in them); one of them treats another person's account of "a Veteran’s difficulty walking" as lay evidence of painful motion for the minimum rating of the joint itself (M21-1 V.iii.1.A.1.f).

A25039304 shows the two side by side. One file held chiropractic notes recording the veteran's statements about his low back pain in relation to knee pain that had "created an antalgic gait" and an August 2024 VA spine examination documenting "muscle spasms resulting in abnormal gait or abnormal spine contour"; the Board listed both as evidence on the link and decided no rating (binding only with regard to the specific case decided, 38 CFR 20.1303).

Two-panel diagram with no people, titled a limp is evidence for the link, not the measure. The left panel is headed knee, and shows a knee icon with a footprint trail labelled limp or altered gait. Its text reads: 38 CFR 4.45, disturbance of locomotion is among the related considerations for the joints. An arrow labelled can be linked to, the medical question under 38 CFR 3.310, runs from the knee heading to the back heading, above the two rows. The right panel is headed back, General Rating Formula for the spine, and shows two rows. The first reads: one option in the 20 percent row, muscle spasm or guarding severe enough to result in an abnormal gait or abnormal spinal contour. The second reads: one option in the 10 percent row, muscle spasm, guarding, or localized tenderness not resulting in abnormal gait or abnormal spinal contour. A tag beside the rows reads: a limp from the knee is not spasm or guarding in the back. A bar across the bottom reads: 38 CFR 4.14, the evaluation of the same manifestation under different diagnoses is to be avoided. — The limp belongs to the knee's ledger and to the medical question of the link, not to the spine formula.

Can the knee and the back both be rated?

Yes, where the back is service connected and its findings are not the knee's: knee motion and instability are measured at the knee, and spinal motion, spasm, guarding and ankylosis at the spine. 38 CFR 4.14 does not bar that pairing, and VA's manual says "Separate evaluations are warranted where none of the symptomatology for any one condition duplicates or overlaps another" (M21-1 V.ii.3.D.2.b, citing Esteban v. Brown; its Example 3 separates DC 5257 instability from DC 5003 arthritis motion). The back can still be rated 0 percent, and it is combined with the knee, not added.

Where 38 CFR 4.14 bites

38 CFR 4.14 closes with one sentence that holds two bans: "Both the use of manifestations not resulting from service-connected disease or injury in establishing the service-connected evaluation, and the evaluation of the same manifestation under different diagnoses are to be avoided." The word pyramiding appears only in its heading.

The second ban is the limp's: if the limp is part of how the knee is evaluated, it cannot be evaluated again under a spine code. That bars counting it twice; it is no reason to keep it out of the record.

The first ban is about cause: back symptoms from a cause that is not service connected, such as a separate injury, cannot build the service-connected back evaluation. That is not the aggravation baseline: 4.14 governs what may be counted in evaluating a service-connected disability, 3.310(b) how much of a nonservice-connected back's worsening is service connected.

Symptoms that cannot be separated

VA's manual says: "When evaluating an SC disability for rating purposes, if it is not possible to differentiate the Veteran’s symptoms from an SC and NSC disability, all symptoms should be deemed related to the SC disability and rated accordingly" (M21-1 V.ii.3.D.2.c, topic change date July 29, 2026). The block adds that "the NSC disability is not considered SC" and that the separation question "is a medical determination". We found no text on how it works alongside an aggravation baseline, so we do not treat it as settled here.

Document the limp anyway

Dated records of a limp, an antalgic gait, a cane, a brace or a shorter leg are evidence for the link; 18140986 and 24003862, in the Board section below, relied on them (each binding only with regard to the specific case decided). They are simply not scored on the spine. Describe your walking to every clinician and examiner as it is; a record helps only when it is accurate.

The full spine formula, with the rating estimator, is on our back pain page.

What Can Back Pain Secondary to a Knee Pay?

That depends on the route, and either route can pay nothing. These rules are why this page promises no increase.

Causation rates the whole back

Under 38 CFR 3.310(a) the back "shall be considered a part of the original condition", with no baseline deduction, so it is rated on its full current level. For the thoracolumbar spine the formula's levels are 10, 20, 40 and 50 percent, with 100 percent for unfavorable ankylosis of the entire spine; its 30 percent row is for the neck. It prints no 0 percent row, and 38 CFR 4.31 supplies one: "In every instance where the schedule does not provide a zero percent evaluation for a diagnostic code, a zero percent evaluation shall be assigned when the requirements for a compensable evaluation are not met." A back whose findings fall short of the 10 percent row is rated 0 percent and is still service connected, unless painful motion is shown. Applying 38 CFR 4.59, VA's manual says the minimum compensable evaluation "must be assigned when painful motion is demonstrated" and is the lowest evaluation the code specifies, 10 percent on the spine formula (M21-1 V.iii.1.A.1.g); "Lay evidence or subjective reporting of painful motion is sufficient, if found credible" (V.iii.1.A.1.f).

Aggravation rates only the increase

Only the worsening is service connected. In the words of Block 1.g, "The evaluation assigned will reflect the difference between the current and baseline levels of severity." A back now at the 20 percent level over a baseline at the 10 percent level is rated 10 percent.

Block 1.c accepts three kinds of evidence for the baseline: medical evidence from before the onset of aggravation; the earliest medical evidence between that onset and the evidence of current severity; and "any other credible evidence, including lay evidence, that supports establishment of the baseline", though lay evidence alone may fall short. Setting it "is a legal determination made by the rating activity" (Block 1.d): a letter can supply evidence about the baseline but cannot set it, and a letter written today is not a record from before the worsening.

"In accordance with 38 CFR 3.310(b), if no baseline can be established, no aggravation can be demonstrated" (Block 1.f). Its note, that VA should "not assume a baseline of 0 percent when a baseline cannot be established by the available evidence", works against the veteran: the award is the current level minus the baseline, so VA may not treat a missing baseline as zero and pay the full current level. The protection is that a denial for lack of a baseline comes only "after VA’s duty to assist has been fulfilled". Permanent worsening is not required (Block 1.a).

Same level means 0 percent

Block 1.h: "Whenever there is competent and credible evidence of an increase in severity of the disability from the established baseline, but the current and baseline disability would be assigned the same level of evaluation under the rating schedule, grant SC and assign a noncompensable evaluation." It goes on: "To justify SC, the degree of disability after aggravation does not have to be at least one level of evaluation higher than the baseline." So a back at the 20 percent level before and after, with some worsening shown, is service connected at 0 percent.

Two ratings are combined, not added

38 CFR 4.25 combines ratings through Table I, greatest disability first, then rounds: "This combined value will then be converted to the nearest number divisible by 10, and combined values ending in 5 will be adjusted upward." The conversion "will be done only once per rating decision, will follow the combining of all disabilities, and will be the last procedure in determining the combined degree of disability" (4.25(b)). Its headnote reads "10 combined with 10 is 19".

38 CFR 4.25 Table I, where these are your only two ratings

38 CFR 4.25 Table I, where these are your only two ratings
Your other rating	Back rating	Table I value	Combined after rounding
10 percent	20 percent	28	30 percent
10 percent	10 percent	19	20 percent
20 percent	20 percent	36	40 percent
30 percent	10 percent	37	40 percent
40 percent	20 percent	52	50 percent
60 percent	10 percent	64	60 percent (no change)
70 percent	10 percent	73	70 percent (no change)

Cells read from Table I in 38 CFR 4.25 (eCFR issue of August 27, 2026). Rounding happens once, after every rating is combined, so with other ratings in the file these rows do not carry over.

Two-column diagram with no people, titled what a back grant can pay, where a knee rated 10 percent and the back are your only two ratings. The left column is headed causation, 38 CFR 3.310(a), and reads: the back is considered a part of the original condition and is rated on its current level. Its example: back now at the 20 percent level, back rating 20 percent; 38 CFR 4.25 Table I, 20 and 10 give 28, rounded to 30. The right column is headed aggravation, 38 CFR 3.310(b), and reads: only the increase over a documented baseline is rated. Its first example: baseline at the 10 percent level, back now at the 20 percent level, back rating 10 percent; 10 and 10 give 19, rounded to 20. Its second example: baseline and now both at the 20 percent level, with some worsening shown, service connection granted at 0 percent, noncompensable, M21-1 Part V, Subpart ii, Chapter 2, Section D, Block 1.h. A footer reads: if no baseline can be established, no aggravation can be demonstrated, Block 1.f; rounding to the nearest 10 happens once, after every rating is combined, so with other ratings in the file these rows do not carry over. — Two routes to the same back, and what each can pay where the knee and the back are your only two ratings.

How the spine formula measures flexion, combined motion and ankylosis, and the estimator that applies it, are on our back pain page.

What About Sciatica or Nerve Pain From Your Back?

Nerve pain down a leg from the back belongs to the back claim, not to a new link with the knee.

Note (1): Evaluate any associated objective neurologic abnormalities, including, but not limited to, bowel or bladder impairment, separately, under an appropriate diagnostic code.38 CFR 4.71a, General Rating Formula for Diseases and Injuries of the Spine

If an evaluation is assigned based on incapacitating episodes, a separate evaluation may not be assigned for LOM, radiculopathy, or any other associated objective neurological abnormality as doing so would constitute pyramiding under 38 CFR 4.14.M21-1 V.iii.1.B.3.a (topic change date February 27, 2025)

For that reason and because neurological complications of spinal disease are contemplated in the evaluation criteria for spinal conditions under 38 CFR 4.71a, a claim asserting new complications of spinal disease is considered a claim for increase rather than a claim for secondary SC.M21-1 V.iii.1.B.3.d (topic change date February 27, 2025)

Rated with the back, separately

VA's manual evaluates spine conditions "by assigning separate evaluations for the orthopedic and neurological manifestations" (M21-1 V.iii.1.B.3.a), subject to the exception below. Where a service-connected thoracolumbar disability is present and radiculopathy is diagnosed without a specific nerve root identified, the manual rates the leg under the sciatic nerve, 38 CFR 4.124a, DC 8520 (V.iii.1.B.3.d). A separate nerve evaluation is combined with the back rating under 38 CFR 4.25, not added, so it can leave the combined rating unchanged. Where each leg carries a compensable rating, such as the knee on one side and a sciatic nerve rating on the other, 38 CFR 4.26 applies the bilateral factor first: the two legs' ratings are combined and "10 percent of this value will be added (i.e., not combined)" before further combining, and 4.26(a) covers paired legs "regardless of location or specified type of impairment".

The one exception: incapacitating episodes

DC 5243 reads: "Assign this diagnostic code only when there is disc herniation with compression and/or irritation of the adjacent nerve root; assign diagnostic code 5242 for all other disc diagnoses." IVDS is evaluated under the General Rating Formula or the incapacitating-episodes formula, "whichever method results in the higher evaluation when all disabilities are combined under § 4.25." That formula rates the total duration of episodes in the past 12 months at 10, 20, 40 or 60 percent, from at least one week to at least 6 weeks.

An incapacitating episode "requires bed rest prescribed by a physician and treatment by a physician" (IVDS Note (1)); in the manual's words, "By definition, an incapacitating episode of IVDS requires bedrest prescribed by a physician" (V.iii.1.B.3.c). A back rated that way carries no separate nerve evaluation (the B.3.a sentence above, which B.3.d names as the exception); the CFR text itself does not state it.

A new nerve problem is a claim for increase

Neurological complications "represent a medical progression or worsening of the spinal disease" (V.iii.1.B.3.d), so sciatica that appears after the back is service connected is part of the back claim, and when a spinal disability is expressly claimed, VA must consider its neurological complications as within scope of the claim. B.3.d points to the effective-date provisions for increases, and its references list 38 CFR 3.400(o), whose paragraph (o)(2) begins: "Earliest date as of which it is factually ascertainable based on all evidence of record that an increase in disability had occurred if a complete claim or intent to file a claim is received within 1 year from such date, otherwise, date of receipt of claim."

Can a Bad Knee Cause Back Pain? What the Research Shows

No study we found shows that knee problems cause back problems in a population. A plausible mechanism exists for one pathway, the research is limited, and it runs in both directions. Who was studied matters. In the knee-to-back studies here the knee condition was arthritis or a knee replacement, apart from Young's kneecap pain and Murata's 366 patients, whose condition the abstract does not state; the brace and induced-limp experiments used healthy volunteers. Two of the knee replacement series excluded arthritis after an injury (Kechagias 2024; Londhe 2021; both full text), and none of the studies here followed a ligament or meniscus tear to later back outcomes. Figures are from abstracts unless marked full text.

Knee extension loss: lab and model studies

The best-supported pathway, though none of these studies measured back pain. A brace that kept 10 healthy women from straightening one knee by the last 30 degrees changed trunk and pelvic motion; a 15-degree block did not significantly change trunk motion in that study (Harato 2008, PMID 18585042). In 17 healthy young men, braced loss of knee extension raised computer-estimated L4-5 forces in walking, forward by up to 36.5 percent and vertical by up to 23.0 percent, more with more restriction; at 15 degrees the forward force and the forward lean of the trunk rose too, while the vertical force rose significantly only at 30 degrees (full text) (Nakatsuji 2022, PMID 36046010). In 17 adults with knee arthritis examined once (9 with an extension deficit of 10 degrees or more, 8 with less; full text), a walking model estimated higher short peaks of upper-lumbar disc compression in the deficit group, while average compression did not differ; that group also had a different spine and pelvis shape on x-ray (smaller pelvic incidence, larger sagittal vertical axis) (Kotaki 2025, PMID 41228828). In 366 patients examined once, more than 5 degrees of lost knee extension went with less lumbar curve on a standing x-ray; age may confound both, and the abstract reports no symptom measure (Murata 2003, PMID 12585585). No contracture threshold follows from these.

Trunk lean: very low certainty

A 2019 review that included 32 studies (2,037 people), pooling data where possible, found that people with knee arthritis leaned sideways toward the affected (ipsilateral) leg more than healthy adults while walking (pooled standardized mean difference 1.18, 95 percent confidence interval 0.59 to 1.77), at a certainty the authors graded "very low"; forward lean was not significant, and back pain was not measured (Iijima 2019, PMID 30904334).

A limp induced in healthy volunteers: little change in modeled walking loads

In 12 healthy adults made to walk unevenly, modeled low back demands were "similar to, or only moderately different from normal walking for most conditions" (the authors' words; Banks 2022, PMID 36095916). With a walking boot and carried loads, peak forces were "mostly not impacted" (again theirs) except with a 15 percent body-weight load carried alone in the hand opposite the boot (Banks 2026, PMID 42608014). Neither tested a limp from a knee condition, and the authors pointed to daily tasks other than walking, and to factors other than mechanics, as what to study next.

Leg length: mixed

In 148 Finnish food-industry workers, a leg-length difference of 6 mm or more went with more intense low back pain, in standing but not seated workers, a difference the study did not confirm statistically (full text); past leg fractures were excluded (full text), so none of these differences came from a broken leg (Rannisto 2015, PMID 25943907). A 2025 review of observational studies comparing people with and without low back pain found "leg length discrepancy showed no difference between groups" (Sugavanam 2025, PMID 39166267). Neither concerned a shortening caused by a knee condition.

Knee and back trouble together: co-occurrence, not order

Knee and lower-spine arthritis on x-ray occur together more than chance predicts (pooled odds about 1.8), but from unadjusted, cross-sectional data (full text), so neither age-adjusted nor time-ordered (Betzler 2022, PMID 35421911). Of 9,057 US knee replacement patients, 52.6 percent reported back pain at surgery (Ayers 2023, PMID 37001625). In 205 people with knee arthritis, varus thrust, a visible outward lurch of the knee in walking, went with higher odds of moderate or severe back pain at the same visit (Iijima 2020, PMID 31254449).

Which came first: the time-order studies

In a North Carolina community study of 819 adults with paired spine x-rays, followed for an average of five and a half years (full text), x-ray knee arthritis, adjusted for age, weight and other joints, did not significantly predict new low back pain (about 500 people analyzed) or worsening low back pain (about 650) (full text), results that do not rule a link out; it did predict new lumbar facet arthritis (hazard ratio 4.18, 95 percent confidence interval 1.44 to 12.2), an x-ray finding with a wide margin (Goode 2022, PMID 33973412). In a Swedish register of 548,681 people, new knee or hip arthritis went with more later diagnoses of back pain, depression, cardiovascular diseases and osteoporosis (adjusted hazard ratios 1.07 to 1.60 across these outcomes), with no knee-only figure in the abstract (Dell'Isola 2022, PMID 34086422). Of 92,319 Military Health System patients aged 17 to 60 with kneecap (patellofemoral) pain, 21.2 percent were diagnosed with a lumbar injury within two years; there was no comparison group without kneecap pain, and the authors say causation cannot be determined (Young 2023, PMID 36793225).

The reverse order

Studies have also found the reverse order: in 4,638 older community residents followed about five years, back pain came before new knee pain (Ito 2019, PMID 30987680), and a genetic study linked predicted disc degeneration to knee arthritis (Zhu 2026, PMID 42770026), so an association cannot by itself show which caused which.

Shared risk factors an examiner will raise

VA's manual: "In degenerative arthritis, designated for rating purposes under 38 CFR 4.71a, DC 5003, age is the most powerful risk factor" (M21-1 V.iii.1.C.3.a; about DC 5003 arthritis in general, not a lumbar finding; "most" is the manual's word). The VA/DoD low back pain guideline (2022), which says it is "not intended to represent Department of Veterans Affairs or TRICARE policy", puts degenerative lumbar stenosis at "20% in <40 years and 47.2% in >60 years"; its source, a CT study of 191 Framingham participants, gives 47.2 percent for ages 60 to 69, counting acquired narrowing of the spinal canal at a 12 mm cut-point (Kalichman 2009, PMID 19398386). Disc degeneration appears on spine imaging in 37 percent of 20-year-olds and 96 percent of 80-year-olds without symptoms (Brinjikji 2015, PMID 25430861), though it is more prevalent in adults 50 or younger with back pain (odds ratio 2.24, 1.21 to 4.15; Brinjikji 2015, PMID 26359154). Obesity predicted new low back pain (hazard ratio 1.75, 1.19 to 2.56; Goode 2022); a 2023 review of 13 studies reported high pelvic incidence as a risk factor for both spondylolisthesis and knee arthritis, a finding it drew from one included cohort (full text) (Amarasinghe 2023, PMID 36864486); arthritis and spinal stenosis share genetic risk (genetic correlation 0.47 to 0.66), though the same study's genetic analysis read arthritis at any joint, not the knee alone, as a contributor to spinal stenosis (odds ratio 1.6, 1.41 to 1.79) rather than only a shared risk (full text) (Sobczyk 2024, PMID 38160745); and genetic estimates for knee arthritis and back pain are very small and conflict (Gong 2024, PMID 39678046). For scale, the guideline cites a national survey in which 33 percent of Veterans reported significant back pain in the prior three months, and an annual low back pain incidence of 12.0 percent among active duty Service Members from 2010 to 2014.

After knee replacement: uncontrolled series

In 110 patients having knee replacement for knees that could not straighten, 25 of the 73 patients with back pain beforehand (full text) reported less or none a year later, with the lumbar curve not significantly changed (Kitagawa 2021, PMID 32693445). Of 45 patients, 28 had moderate or worse back pain before surgery and about half of them reported less at 12 months, with no comparison group and arthritis caused by trauma excluded (full text) (Kechagias 2024, PMID 38716012). In another series 108 of 120 improved on the Oswestry index, which knee pain may also move, and a letter asked how that was controlled (Londhe 2021, PMID 33355851; letter, PMID 35508940). Pooled data show pelvic tilt changed significantly and lumbar lordosis did not change significantly (Jain 2025, PMID 40698278). Relief after knee surgery does not show what caused the back pain.

What the research supports

A plausible mechanism for loss of knee extension, from studies without a back pain outcome.
A sideways trunk lean in knee arthritis, at very low certainty.
Little change in modeled walking loads when healthy volunteers were made to walk unevenly, and mixed results for leg length.
Co-occurrence without an order, and shared risk factors an examiner can point to.

What it does not support

Any population figure for the share of back conditions a knee condition causes; we found none.
A settled answer either way: the time-order and genetic studies are mixed, so ruling the link out overstates the research as much as ruling it in.
A contracture threshold, a leg-length size that causes back pain, or relief after surgery as evidence of cause.

So the research makes the question fair to weigh in one veteran's file without settling it, and in the Board decisions we read, the weight fell on that file: the gait history, the time order, the examination, and reasoning that addresses age, weight and other causes (each binding only with regard to the specific case decided, 38 CFR 20.1303).

What Board Decisions on Back Claims Secondary to a Knee Show

We read 80 Board decision files and rely on 16 in which a back claim ran through a knee, alone or with other joints (in one, 24001968, the knees themselves were denied), each read at its ORDER block and the analysis behind it. Five granted service connection for the back and none set a percentage; two denied it; nine sent it back for a new or better opinion. Four of the five grants rested on private opinions, one from a nurse practitioner. No decision in the set applied the 38 CFR 3.310(b) baseline or cited VA's manual arthritis rule. The newest is from April 2025, so every one predates the manual's May 1, 2026 update; the eight from 2018 also predate its arthritis topic's February 8, 2021 change date, and none of the 16 applied a diagnostic code. Seven come from web searches of the Board's archive and nine were carried over from earlier research, so the set is neither complete nor random, and these counts are not grant rates and predict nothing.

The rule is 38 CFR 20.1303: previously issued Board decisions "will be considered binding only with regard to the specific case decided", and "each case presented to the Board will be decided on the basis of the individual facts of the case in light of applicable procedure and substantive law." Every entry below is described for the issue it decided and nothing more.

A grant of service connection is not a grant of a percentage. The rating comes afterward, under the spine formula or, on aggravation, as the increase over a baseline, and none of these decisions shows how the back was then rated. Three grants do not say which route they took: two found the back "caused or aggravated" or "caused and/or aggravated", and 21066807 found it "related to a service-connected disability".

Service connection granted

Citation Nr 24003862 (decided January 25, 2024)

Granted: a back disability secondary to the right knee, after reopening; no percentage. Legacy docket, after Spicer.

An October 2017 private nurse practitioner's letter (NP G.G.U.) cited journal evidence linking low back pain to antalgic gait problems due to knee issues and a January 2017 note of an antalgic gait; the Board found the back (lumbar degenerative disc disease and a herniated disc) "proximately caused and/or aggravated" by the knee. Three VA opinions lost because they "do not consider the Veteran's antalgic gait due to his right knee condition".

Why it matters. The same author's sleep apnea opinion lost because it "does not rest on correct facts".