OPINION
PAUL W. GREEN, Justice.Edward and Karen Lette (“the Lettes”) appeal the no-evidence summary judgment rendered against them in their medical malpractice law suit. The summary judgment was based on the trial court’s conclusion that the Lettes had introduced no evidence to support the causation element of their claim. The Lettes assert the trial court erred in granting the summary judgment because it applied an erroneous standard of proof and because there was evidence of causation. We affirm.
Background
Edward Lette underwent prostate surgery on October 1, 1998. At 9:00 a.m. on October 2nd, he was given 100 milligrams of the blood anticoagulant, Lovenox. Another 100 milligrams of Lovenox was administered at 10:00 p.m. Although Mr. Lette was receiving pain medication through a morphine pump, his post-surgery pain was unrelieved. At 2:05 a.m. on October 3rd, in an effort to relieve his pain, Mr. Lette was given 25 milligrams of Toradol, a non-steroidal anti-inflammatory agent that also has the effect of impairing blood platelet aggregation, which is part of the blood clotting process. The Toradol was administered contrary to the surgeon’s orders, but was consistent with the anesthesiologist’s orders. Later that same day, Mr. Lette was required to undergo a second surgery to remove a hematoma, or blood clot, caused by arterial bleeding in his pelvic area. Mr. Lette subsequently developed a condition known as complex regional pain syndrome which was associated with the hematoma. Mr. Lette claims the arterial bleeding and resulting injury were caused by the improper post-surgery administration of the Toradol, which he says prevented his blood from properly clotting.
Standard of Review
We apply the usual standard of review for no-evidence summary judgments, viewing the evidence in the light most favorable to the respondent against whom the summary judgment was rendered. Moore v. K Mart Corp., 981 S.W.2d 266, 269 (Tex.App.-San Antonio 1998, pet. denied).
Proximate Causation
To establish proximate cause, a plaintiff must prove foreseeability and cause-in-fact. Travis v. City of Mesquite, 830 S.W.2d 94, 98 (Tex.1992). Cause-in-fact is established by showing that the negligence is a substantial factor in bringing about the harm and without which the harm would not have occurred. Kramer v. Lewisville Mem’l Hosp., 858 S.W.2d 397, 400 (Tex.1993). In other words, it must be shown that “but for” the negligence, the injury would not have occurred. Greene v. Thiet, 846 S.W.2d 26, 30 (Tex.App.-San Antonio 1992, writ denied). In medical malpractice cases, the causal connection between the negligence and the injury must be based on expert testimony establishing “reasonable medical probability,” not-mere conjecture, speculation, or possibility. Park Place Hosp. v. Estate of Milo, 909 S.W.2d 508, 511 (Tex.1995). More*602over, other possible causes of the injury must be ruled out with reasonable certainty. Helm v. Swan, 61 S.W.3d 493, 497-98 (Tex.App.-San Antonio 2001, pet. denied).
Discussion
The question is whether there is probative medical evidence that Mr. Lette’s injury would not have occurred but for the administration of the Toradol. There is no direct evidence on this point; that is, there is no expert testimony that, within a reasonable degree of medical probability, the Toradol caused Mr. Lette’s arterial bleeding. Indeed, the medical testimony pointedly avoids making that conclusion.
The most favorable medical testimony comes from Dr. Joel Moake, a hematologist, who testified that when the Toradol was administered to Mr. Lette, “it was not helpful and it probably potentiated, either caused or potentiated bleeding.” 1 But he was unable to say whether the Toradol actually precipitated Mr. Lette’s arterial bleeding or only made it harder for the blood to clot. And regarding the latter possibility, Dr. Moake said Mr. Lette’s ability to clot was already somewhat impaired by the Lovenox that remained in his system when the Toradol was introduced.2 Accordingly, he was unable to quantify what additional effect the Toradol would have had on Mr. Lette’s ability to clot. Other expert testimony offered nothing more favorable to the Lettes on the causation issue.
The Lettes contend, however, that causation may be supported by circumstantial evidence and factual inferences that can reasonably be drawn from the evidence. They contend that, when considering the time-line of events and the testimony regarding the therapeutic effects of the various drugs administered to Mr. Lette, it is more likely than not that the Toradol caused Mr. Lette’s arterial bleeding. As superficially compelling as this argument is, it nonetheless fails to satisfy the well-established standard for proof of causation in medical negligence cases.
The Lettes’ own medical experts were unable to render an unequivocal opinion that the Toradol caused Mr. Lette’s arterial bleeding and resulting injury. Consequently, any conclusions drawn from the time-line of events and testimony concerning the therapeutic effects of Toradol would necessarily be based on mere conjecture, speculation and possibility, a standard of proof that is far below that which is permitted by law. Accordingly, the circumstantial evidence relied on by the Lettes amounts to- no more than a scintilla of evidence of causation and is thus legally insufficient.3
*603Conclusion
After reviewing all the medical evidence tending to favor the Lettes, and disregarding all evidence to the contrary, we find no probative evidence to support the causation element of their claim. Accordingly, the issues are overruled and the judgment of the trial court is affirmed.
Dissenting opinion by PHIL HARDBERGER, Chief Justice.
.The "either/or” choice offered by Dr. Moake fails to quantify the degree to which the administration of the Toradol was related to the arterial bleed; that is, whether the Toradol was a remote, minor or substantial factor in causing the bleeding to occur. Accordingly, the testimony fails to meet the standard required for proof of medical causation. See Sisters of St. Joseph of Texas v. Cheek, 61 S.W.3d 32, 36-37 (Tex.App.-Amarillo 2001, pet. denied) (testimony that acts or omissions "caused or contributed to” death leaves fact finder to speculate on causation and is legally insufficient).
. Because he had an artificial heart valve, Mr. Lette had been on a program of oral Couma-din, an anticoagulant, to prevent blood from clotting around his heart valve. The Couma-din was discontinued a week before the prostate surgery was performed and was resumed on October 3. According to medical testimony, the coumadin could not have been affecting Lette’s blood clotting ability when the arterial bleeding occurred.
. The dissent concludes that "circumstantial evidence was presented in this case, supported by medical literature, from which rea*603sonable minds could draw an inference that the administration of the Toradol was a substantial factor in bringing about Mr. Lette's bleeding complications....” Dissenting op. at -. The problem with this statement is that the doctors who were aware of the time-line of events, and of the medical literature describing the therapeutic effects of the various medications, refused to make the causal connection the dissent says a jury should be allowed to make. It is precisely because a jury is not qualified to draw medical conclusions that expert testimony is required on the element of causation. In fact, with the exception of the negligence per se cases, it is hard to imagine a medical malpractice case where circumstantial evidence alone would be sufficient to raise an issue on causation.