Many aspects of publications on long-term results are imprecise. They often fail to separate metacarpophalangeal and proximal interphalangeal joint contracture which prognosis is very different as was recognised by Honner (1971). They usually classify the results in terms of recurrence and extension of the disease. Gordon (1957) separated these processes by defining disease within the area operated upon as a recurrence and disease beyond the confines of the operative field as an extension. This concept has generally been accepted but demands a more accurate definition. Finally, those studies do not give us informations about the time interval between surgery and the reappearance of clinical signs of activity of the disease. They do not tell us either which factors play a role in that continuing progress of the disease.

This chapter will be devoted to these questions as they apply to segmental aponeurectomy.

Restoration of normal anatomy and function in this area is difficult because any dissection creates new scar tissue which is likely to lead to progressive stiffness and the recurrence of contracture. One of the reasons for this is the imbalance between the strong flexing force of the flexor digitorum superficialis and the weak extension force of the central slip of the extensor apparatus which is responsible of the resting position of the hand in mid-flexion. In addition, this central slip may be attenuated by a prolonged period of flexion.

A progressive recurrence of contracture following surgery for Dupuytren's disease may therefore indicate a mechanical problem rather than a true recurrence, even if the volar tissues are tight. This tightness usually reflects a lack of active extension.

Some authors (Orlando, 1974; Noble, 1976) have based their diagnosis of recurrence by comparing the pre- and postoperative mobility. In the light of the previous discussion, this is probably not correct.

In the following analysis, the diagnosis of recurrence was based on the presence of a nodule or of an identifiable cord without taking the loss of extension into account. The criteria were rather severe since the reappearance of a nodule anywhere in an operated ray was considered as a recurrence even if that precise location was not directly in the original operating field. For example, if a patient was operated for a pretendinous band of the little finger and if on control examination a nodule was found on the ulnar border of the finger in the adductor tendon, the case was coded as showing a recurrence. It was decided to interpret the observations in this way because it is almost impossible to be sure of the exact limits of the original operating dissection.

For these reasons, McGrouther (1990f) favours an analysis of the number of hands which remain clear of disease after a particular operative treatment rather than a study of recurrences and extensions.

In the following analysis, we will try to develop both aspects so that a comparison with other studies, though difficult, remains possible.

In fact in hands that did not show signs of extension or recurrence, the skin was supple and no induration that would give evidence of the presence of residual pathological tissue was palpable.

We were able to review 173 (59.2 %) of the 292 segmental aponeurectomies for which a control at least one year after the operation was possible. This represents 141 patients (58.8 %) on 240. The patients who came in the first year after the operation with signs of progress of the disease are included in this review (Moermans, 1996).

Figure 14-1: Interval between operation and follow-up examination |

Figure 14-2: Observed continuing activity of the disease |

In most operated rays that did not show signs of recurrence, the tissues were supple and the segments of diseased fascia that were left in place at the time of the operation had softened and completely disappeared.

Recurrences were not evenly distributed among the rays, being much more common in the fifth ray (table 1).

Table 2 shows the number of recurrences in relation to the number of contractures originally operated upon. As can be seen the proportion of recurrence was much higher in the radial rays and in the first web than in the other rays. The percentage is also twice as high in the little finger than in the central rays. Those differences are statistically highly significant (p<0.01; statistical methodology: Chi-square test).

If we split those numbers by joint level, we see that recurrences were much more frequent at the proximal interphalangeal joint level than at the metacarpophalangeal level (table 3). This difference in the evolution of the palm and the finger is statistically very significant (p<0.00005; statistical methodology: Chi-square test). Moreover, the observed differences between the interphalangeal joints of the five rays are also statistically significant (p<0.05). Between the M.P. joints, they are not (p<0.2).

Figure 14-3: Tubiana's grade |

Figure 14-4: Functional impairment |

The average Tubiana's grade was 1.1 (S.D. 1.5; range 0 to 8) and the average impairment of function was 2.2 (S.D. 3.7; range 0 to 29). The preoperative values for the same hands were respectively 3.0 (S.D. 1.9; range 1 to 12) for the Tubiana's grade and 6.3 (S.D. 5.4; range 1 to 35) for the impairment of function (tables 4 and 5).

Among the 65 recurrences, four (6.2 %) had only a palmar or digital nodule without extension deficit, 23 (35.4 %) had less than 45° of total extension deficit (fig. 3) and 17 (26.2 %) had a functional impairment of 1 % or less (fig. 4). The average Tubiana's grade was 2.3 (S.D. 1.7; range 0 to 8) and the average impairment of function was 4.6 (S.D. 5.0; range 0 to 29).

If we look at the hands which did not show any sign of progression of the disease, we can see that the operation brought a lasting correction of the contracture (table 6).

The follow-up values are even slightly better than the immediate postoperative measurements except for the proximal interphalangeal joint of the little finger.

On these 24 hands which had to be reoperated, 18 had never been operated before the segmental aponeurectomy. Among the other six, 4 had undergone a limited fasciectomy, one a Skoog operation and one a dermo-fasciectomy. This last hand had to be operated for an extension of the disease.

The recurrence rate in this series was 37.6 %, which is comparable to that reported in other studies for other surgical techniques: 34 % in Hakstian's reviews (1966, 1974), 40 % in Hueston's (1961) and 66 % in Tubiana's (1985b). In an exhaustive review of the literature, McGrouther (1990f) found percentages of recurrence oscillating between 2 and 63% and percentages of hands clear of disease varying between 0 and 65% ! Leclercq et al. (1993) reports rates of recurrences between 20 and 68 %. In this review, 46.8 % of the hands had no signs of progression of the disease at the time of the follow-up examination.

For McGrouther those varying percentages of recurrence, extension and hands clear of the disease simply reflect differences in the arbitrary criteria for definition of these terms. There is certainly another valid explanation for these variations: a true comparison of those percentages is impossible because the follow-up periods in the different studies are not the same. This aspect of the problem will be studied in the next chapter.

Nevertheless, the fact that the percentage of recurrences after segmental aponeurectomy does not differ much from that reported in other series indicates that this technique does not create more late problems than with other more extensive procedures. McGrouther (1990f) drew the same conclusion from his own review of late results: "there was no evidence that a more radical operation within the range of procedures performed at Canniesburn Hospital would protect the patient from recurrence. It seems justifiable to perform the simplest possible operation to relieve the contracture and improve the function of the hand".

Another interesting aspect of this review is the stability of the correction of the contracture. For all joints except the proximal interphalangeal joint of the little finger in hands without signs of progress of the disease, the mobility at the time of follow-up was equal and even better than it was in the immediate post-operative period. This is probably due to the limited dissection and hence the limited formation of scar tissue which development on the volar aspect of the finger certainly plays an adverse role.

If we examine the Tubiana's grades and the impairment of function percentages of all reviewed hands, including those with extensions or recurrences, we see that the loss of mobility is limited. The Tubiana's grade goes up from 0.6 to 1.1 and the impairment of function from 0.9 to 2.2 %. This very slow evolution is another good reason for not being too aggressive with the first operation. This is all the more true because only 24 hands had to be reoperated (of these 4 had an extension that would not have been in any case controlled by a more extensive approach).

This review also verified our basic postulate about segmental aponeurectomy. The absence of palpable residual aponeurotic band in the operated hands that did not show signs of recurrence or extension confirms that a cord from which tension has been eliminated disappears or at least ceases to act as a contracture

These complicating factors eliminate the possibility of doing something simple, such as calculating the average time between operation and recurrence. Actually this is what was done in all the above-mentioned studies and with my own material in the preceding chapters (I have done it to have some studies to compare with). The error is that a percentage of recurrence or of patients remaining free of disease was established when all patients were not followed for the same period of time. Patients reviewed after one or two years were included in the computations with patients followed for a much longer period of time even though we do not know when and at which rate recurrences or extensions of the disease really occur.

What we need is a special statistical technique for looking at the interval between two events when the second event does not necessarily happen to everyone and when people are observed for different periods of time.

The basic idea of the life table is to subdivide the period of observation after a starting point, such as an operation, into smaller time intervals, say single years. For each interval, all people who have been observed at least that long are used to calculate the probability of a terminal event, such as a recurrence, occurring in that interval. The probabilities estimated from each of the intervals are then used to estimate the overall probability of the event occurring at different time points. All available data are used for the computations (SPSS^{5}).

From the first line, we see that we have 173 observations entering the study and that 23 of them developed a recurrence in the first year. Initially, we may think that the estimate of the probability of developing a recurrence in the first year should be 23 out of 173. However, such a calculation does not take into account the fact that there are 9 cases who did not complete their first year and were lost for later follow-up (number withdrawn during interval). What we know is that the last time they were seen they had no recurrence. We can assume, for simplicity, that they have been observed, on average, for half of the length of the interval.. Thus each is considered as contributing only half of an observation. So for the first interval, instead of having observations for 173 cases, we have observations for 168.5 (173-(9 x 0.5)) which is the number exposed to risk. From this all other probabilities can be computed. The probability of developing a recurrence in the first year is thus 23/168.5, or 0.1365. The probability of being free of recurrence until the end of the first year is 1 minus the probability of recurrence: 0.8635.

Using the same calculations, we can estimate the risk of developing a recurrence during the second year, assuming that it did not happen during the first year. In this case, it is 0.8566. From these two probabilities (the probability of making it through the first year, and the probability of making it through the second year given that a patient has made it through the first), we can estimate the probability that the patient will make it to the end of the second year. The formula is:

P (second) = P (first) × P (second given first).

In this case, it is: 0.8635 × 0.8566 = 0.7397.

The column headed "Cumulative proportion surviving at end" shows the cumulative proportion of operated cases that did not develop a recurrence before the end of the interval. The first row contains the data pertaining to the first year of observation. The starting point is the date of the operation. Figure 5 shows a graph of the cumulative proportion of cases that did not develop a recurrence.

Figure 14-5: Cumulative proportion of cases that did not develop a recurrence |

What we observe from the table and the graph is that the proportion of cases that develop a recurrence is not constant with time. It seems to be almost constant during the first four years but then drops to become zero in the last years. The cumulative proportion of cases that did not develop a recurrence seems to level off at 44 % though the proportion for the last three intervals is calculated on a small number of cases.

The introduction by the life table analysis of corrective factors allowing for the different follow-up periods brings about, as expected, a much higher recurrence rate than I had calculated with the raw data: 56 % compared with 37.6 %. This proportion is much closer to the 66 % reported by Tubiana and Leclercq (1985) in their late review of 50 hands operated between 8 and 14 years before (average 10 years).

Figure 14-6: Proportion of cases that developed a recurrence |

To define the

*R ^{2}* may be interpreted as the proportion of the total variation of the dependant variable around its mean that is explained by the model. The

Figure 14-7: Predicted and observed proportion of recurrences |

From the

Figure 14-8: Comparison of survival functions for cases with and without a family history of the disease |

Figure 9 shows the survival functions for both sexes.

Figure 14-9: Survival functions for men and women |

Figure 14-10: Survival functions for cases operated for the first time and cases operated previously. |

The median survival time for the 147 cases operated for the first time is 5.9 years compared with 2.8 years for the 26 others. The observed difference is nevertheless not statistically significant.

As can be seen from figure 11, the survival functions are very different for these two groups of cases. This difference is statistically highly significant (p < 0.0001).

Figure 14-11: Comparison of the observed and expected survival functions for cases with and without a total correction of the contracture (Yes = total correction). |

The parameters of the limited growth model were calculated for both groups, as well as the coefficient of determination

The fact that more recurrences occur earlier in cases with a residual contracture probably means that more pathological tissues were left in place during the operation and that the impossibility to achieve a full re-extension was not only due to joint stiffness but also to the presence of residual diseased fascia.

Figure 14-12: Observed and calculated survival functions for cases with and without ectopic localisations of the disease (Yes = presence of ectopic localisations of the disease). |

For cases (61) without ectopic localisations of the disease, the median time without recurrence is superior to 8 years, for cases (28) with ectopic lesions it is equal to 2.3 years. The observed differences between the survival functions is statistically significant (p<0.005).

The parameters of the limited growth model were calculated for both groups. Fifty percent of the cases without ectopic lesions would never develop a recurrence whereas all cases with ectopic lesions would have recurred after 8 years. The coefficients of determination for the calculated curves are equal to 0.99 in both groups.

This confirms, once again, that the presence of ectopic lesions is a good indicator of the severity of the disease.

Figure 14-13: Observed survival functions for patients operated before and after the age of 45 |

Although based on a small number of patients, the observed difference is statistically significant (p<0.005).

Figure 14-14: Observed survival function for alcoholic and non alcoholic patients |

The observed difference is not significant (p<0.3).

Figure 14-15: Observed survival function for epileptic and non epileptic patients |

Figure 14-16: Survival function for cases with and without a history of local trauma |

Figure 14-17: Survival function for cases with more or less than three rays involved |

The introduction of corrective factors allowing for the different follow-up periods has brought about, as expected, a much higher recurrence rate than I had calculated with the raw data: 56 % compared with 37.6 %.

The cumulative proportion of cases which developed a recurrence expressed as a function of the number of years since the operation has the general shape of the restricted growth curve. The estimation of the parameters of this function using non linear regression has shown that this model fits the data particularly well. The predicted proportion of recurrence after 10 years using those parameters is 68%. This percentage is very close to that observed by Tubiana as it was already pointed out before and, in a way, this validates the method.

The comparison of the survival functions of different subgroups of cases has allowed us to isolate several factors that play a role in the rate and the frequency of recurrences.

Three factors are statistically significant in this regard: the incomplete correction of the contracture at the end of the operation, the presence of ectopic sites of the disease and an operation performed before the age of 45. Hueston (1985c, 1990a) has already recognized young patients with knuckle pads as having a stronger diathesis and greater likelihood of recurrence than those patients who are elderly when the condition appears and in whom no knuckle pads are present. Yet he has given no statistical evidence in support of this concept. McFarlane et al. (1990b), has isolated 338 patients who had had no previous operation and who were examined 2 years or more after operation, from his multicentric review. He found that only the factors 'other areas involved' and 'early onset of the disease' have a statistically significant effect upon recurrence and extension when acting alone. He found also that three other factors, 'bilateral disease', 'family history' and 'more than two rays involved', acting alone have no effect on the course of the disease. However, the simultaneous presence of two or three of them would probably be associated with an aggressive disease. In this review, the involvement of more than two rays was not significantly associated with a worse long term prognosis. The associations between several factors have not been tested because the number of cases in each subgroup would have been too small.

From the preoperative evaluation of the patients, we had already concluded that the involvement of ectopic sites by the disease, and particularly the presence of knuckle pads, was a good indicator of the strength of the diathesis. The prognostic value of this sign is confirmed here by its role in the development of recurrences.

The incomplete correction of the contracture at the end of the operation, particularly at the proximal interphalangeal level, is often attributed to joint stiffness and more specifically to the contracture of the check-rein ligaments passing from the proximal edge of the volar plate to the neck of the proximal phalanx. The fact that more recurrences occurred earlier in cases with a residual contracture probably means that pathological tissues were left in place during the operation and that the impossibility to achieve a full re-extension was not only due to joint stiffness but also to the presence of residual diseased fascia. Could this have been avoided? Watson (1979) has recommended a rather aggressive approach to the residual interphalangeal joint contracture and has advised the division of the check-reins. McGrouther (1990e) has reported disappointing results with these soft tissue releases. In segmental aponeurectomy the emphasis is placed on the release of the joint contracture rather than on the total excision of the contracted fascia to avoid unnecessary damages to the soft tissues. A procedure on the joint was never associated with the operation and a persistent residual contracture has always been treated by postoperative splintage.

A more aggressive attitude would perhaps have permitted to avoid some of the observed recurrences but at which price? This is difficult to assess. McFarlane et al. (1990b) found that in patients who underwent an interphalangeal joint procedure, the complications were twice as frequent and that they were more likely to need postoperative therapy. Of course, those patients had a more aggressive disease. A clear description of the risks associated with aggressive procedures in the digits has been given in the chapter devoted to the early results. If we compare the results on the P.I.P. joint of the little finger reported by McFarlane for conventional procedures with those obtained with segmental aponeurectomy, the difference in favour of the latter is evident: 20 % of perfect results in McFarlane's report compared with 47 % in mine, 25 % of 'same or worse' compared with 3.6 %. The remote possibility of avoiding a few recurrences or more probably of delaying them (remote because overall there is no argument indicating a higher recurrence rate after segmental aponeurectomy) does not warrant such big differences in the postoperative results.

Alcoholism and epilepsy are usually associated with a higher recurrence rate (Hueston, 1990a; McFarlane et al., 1990). It does not appear from this review that these factors play a significant role.

- the operation brought a lasting correction of the contracture in the hands which did not show any sign of progression of the disease; the follow-up values are even slightly better than the immediate postoperative measurements except for the proximal interphalangeal joint of the little finger;

- the proportions of recurrences, extensions and hands free of the disease are similar to those published after other classical procedures;

- most studies are invalidated by the fact that the follow-up periods are unequal;

- the introduction, by life table analysis, of corrective factors allowing for the different follow-up periods gives, by projection, a proportion of 68 % of recurrences after ten years; this percentage is very close to that published by Tubiana, for classical procedures, in a very long term review;

- this very high percentage of recurrences to which a number of extensions should be added confirms that surgery is not curative in Dupuytren's disease; after ten years, almost no hand would be left clear of the disease;

- as far as comparisons are possible, the type of operative approach does not make a clear difference as to the progression of Dupuytren's disease;

- patients operated before the age of 45 or having ectopic sites of the disease run a higher risk of recurrence;

- no other factors, local or general, plays a prognostic role.