In this medical malpractice action, plaintiffs appeal an order denying a motion for new trial. The sole issue on appeal is whether the trial court erred in giving the jury an instruction which deviated from the Standard Jury Instructions by adding the statement that "no physician can be required to guarantee results”. Plaintiffs’ attorney made a timely objection to the instruction on the record and also raised the issue in their motion for new trial. The record discloses that the issue of guaranteed results was never *243injected into the trial by evidence or arguments. We hold that such an instruction constituted reversible error and we remand for a new trial. In so doing, we create a conflict with another panel of this Court which held to the contrary. Warfield v Wyandotte, 117 Mich App 83; 323 NW2d 603 (1982), lv den 417 Mich 919 (1983). Accordingly, an order will be entered simultaneously with this opinion certifying this case to the Michigan Supreme Court for resolution of the conflict.

The trial court correctly gave SJI 30.01, and then added the following language:

"No physician can be required to guarantee results, but the law demands that they bring and apply to the case at hand that degree of skill and care, knowledge and attention ordinarily possessed and exercised by other orthopedic surgeons in the same specialty under like circumstances.”

This addition to the Standard Jury Instructions was taken directly from Warfield, supra, and the trial judge cannot be faulted for so doing. Warñeld had been decided at the time the instruction was given and the trial judge did not commit error in following that case.

However, we believe that Warñeld was wrongly decided since it conflicts with Javis v Ypsilanti Bd of Ed, 393 Mich 689; 227 NW2d 543 (1975). Javis holds that where there is a deviation from an accurate jury instruction prejudicial error will be presumed provided that the deviation was brought to the attention of the trial court prior to commencement of jury deliberations. We would agree with the rationale of Warñeld where no objection is made to the instruction. The addition to the Standard Jury Instructions appears relatively harmless on its face and is arguably nothing more *244than an attempt by the court to clarify the law by contrasting to the jury the difference between what the law is and what it is not. On the surface no harm seems to have been done.

However, a new issue is injected into the case without explanation, without argument, and without any evidence appearing in the record to justify it. We do not know whether any jurors would have been influenced by this remark and, since we do not know, we must be guided by the presumption of reversible error stated in Javis, supra. Since it is not possible for anyone to know which factors will most influence a jury, we have little choice but to follow the guidelines of the Standard Jury Instructions when they are requested by a party and we must strictly conform to their language and form when a party objects to a deviation, especially when the deviation has already been held to be erroneous when it stands alone, as was the instant deviation in Cleveland v Rizzo, 99 Mich App 682; 298 NW2d 617 (1980), lv den 411 Mich 884 (1981).

Reversed and remanded for a new trial with instructions to the trial judge to hold this case in abeyance pending resolution by the Michigan Supreme Court of the conflict between this case and Warfield, supra.

Costs to abide the final outcome.

(dissenting). I respectfully dissent.

At the time of trial, the trial judge was entitled, if not obligated, under Warfield v Wyandotte, 117 Mich App 83; 323 NW2d 603 (1982), lv den 417 Mich 919 (1983), to give the instruction complained of. At the very least, it was not then error to give it.

While my personal predilection is that the instruction complained of is argumentative and *245should not be given except possibly where special circumstances inject such an issue into the case, that is not relevant here. This case was cleanly tried, and the jury found no cause of action. As indicated, there was no error unless we depart from Warñeld retroactively. But, even if we assume it was error, I would believe it to be an insignificant and harmless error which did not affect the result.

I would vote to affirm.